A Fusion Steganographic Algorithm Based on Faster R-CNN

Size: px
Start display at page:

Download "A Fusion Steganographic Algorithm Based on Faster R-CNN"

Transcription

1 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 A Fuson Steganographc Algorthm Based on Faster R-CNN Ruohan Meng 1, 2, Steven G. Rce 3, Jn Wang 4 1, 2, * and Xngmng Sun Abstract: The am of nformaton hdng s to embed the secret message n a normal cover meda such as mage, vdeo, voce or text, and then the secret message s transmtted through the transmsson of the cover meda. The secret message should not be damaged on the process of the cover meda. In order to ensure the nvsblty of secret message, complex texture objects should be chosen for embeddng nformaton. In ths paper, an approach whch corresponds multple steganographc algorthms to complex texture objects was presented for hdng secret message. Frstly, complex texture regons are selected based on a knd of objects detecton algorthm. Secondly, three dfferent steganographc methods were used to hde secret message nto the selected block regon. Expermental results show that the approach enhances the securty and robustness. Keywords: Faster R-CNN, fuson steganography, object detecton, CNNs, nformaton hdng. 1 Introducton Wth the rapd development of the nternet and nformaton technology, how to protect the transmsson of mportant nformaton has ganed a lot of attentons. Informaton hdng s an approach of puttng the secret nformaton nto a carrer (such as dgtal mages) from sender to recever. The recever extracts and restores the orgnal embedded secret nformaton through a specfc method. The carrer that carryng secret nformaton s called cover, whch has certanly sgnfcance n tself. For example, t can be an mage, a document, etc. The carrer adds secret nformaton called stego. In the deal stuaton, stego does not arouse suspcon by attacker n the dssemnaton process. Accordng to the dfferent use of nformaton hdng, t s dvded nto steganography and dgtal watermarkng, the former s manly used for the transmsson of secret nformaton, and the latter s manly used for the protecton of ntellectual property. Accordng to the technology of nformaton hdng, t can be dvded nto a varety of steganographc modes n spatal doman, transform doman and compressed doman. In the early days, the 1 School of Computer and Software, Nanjng Unversty of Informaton Scence and Technology, Nng Lu Road, No. 219, Nanjng, , Chna. 2 Jangsu Engneerng Centre of Network Montorng, Nng Lu Road, No. 219, Nanjng, , Chna. 3 Department of Mathematcs and Computer Scence, Northeastern State Unversty Tahlequah, OK 74464, USA. 4 School of Computer & Communcaton Engneerng, Changsha Unversty of Scence & Technology, Changsha, , Chna. * Correspondng author: Xngmng Sun. Emal: sunnudt@163.com. CMC. do: /cmc

2 2 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 smplest and most representatve method of nformaton hdng n space doman s to hde the nformaton by usng the least sgnfcant bt (LSB) of the mage or all bt algorthms of multple bt planes. However, these algorthms are not robust enough to keep statstcal characterstcs, so that attackers can accurately estmate the embeddng length accordng to the statstcal detecton methods. The typcal detecton algorthms are RS (regular and sngular) groups method [Frdrch and Goljan (2002)], SPA (sample par analyss) method [Dumtrescu, Wu and Wang (2003)] and JPEG compatblty analyss [Frdrch, Goljan and Du (2001)] and so on. Wth the contnuous development of steganography, the newly proposed steganography algorthm can mantan more complex mage statstcal features. For example, HUGO [Pevný, Fller and Bas (2010)], WOW [Holub and Frdrch (2012)], SUNIWARD [Holub, Frdrch and Denemark (2014)] and other content adaptve steganographc algorthms proposed n recent years can automatcally embed the secret nformaton nto the rch nosy texture area on the cover mage to mantan hgh-level statstcs. Deep learnng s well known as a revoluton n machne learnng, especally n the feld of computer vson. In tradtonal approaches of mage feature extracton, the features of SIFT and BOW are usually used as representaton of mages. In recent years, the applcatons of these features extracton nclude mage retreval, mage forenscs and other prvacy protecton felds. Xa et al. [Xa, Zhu, Sun et al. (2018); Xa, Xong, Vaslakos et al. (2017)] apply mage retreval to cloud computng. Zhou et al. [Zhou, Yang, Chen et al. (2016); Zhou, Wu, Huang et al. (2017); Zhou, Wu, Yang et al. (2017); Zhou, Wang, Wu et al. (2017); Cao, Zhou, Sun et al. (2018); Zhou, Mu and Wu (2018)] proposed to apply the tradtonal features of the mage to coverless nformaton hdng. Yuan et al. [ Yuan, L, Wu et al. (2017)] use CNN to detect fngerprnt lveness. However, the appearance of mage depth learnng features makes the feature extracton more rapd and accurate. In the feld of mage classfcaton, deep convolutonal neural networks (DCNN) such as VGGNet [Smonyan and Zsserman (2014)], GoogLeNet [Szegedy, Lu, Ja et al. (2015)] and AlexNet [Krzhevsky, Sutskever and Hnton (2012)] are performance excellent. Gurusamy et al. [Gurusamy and Subramanam (2017)] used a machne learnng approach for MRI Bran Tumor Classfcaton. Based on the achevements above, object detecton has been rapdly developed that detects semantc objects of a class (such as a dog, vehcle, or person) n dgtal mages and vdeo. For texture-rch target mages, the appearance features of object nstances are manly dentfed and detected by extractng stable and abundant feature ponts and correspondng feature descrptors. In deep learnng fles, Faster R-CNN [Ren, Grshck, Grshck et al. (2017)], R-FCN [Da, L, He et al. (2016)] and SSD [Lu, Anguelov, Erhan et al. (2016)] three object detecton models that most wdely used currently. Amng at nformaton hdng, steganalyss approaches based on deep learnng has appeared one after another. These methods have a very good test for the current steganography algorthm. For example, Ye et al. [Ye, N and Y (2017)] proposed to mprove the CNN model to detect hdden nformaton has reached 99 % of the detecton rate. At the same tme, some researchers turned ther attenton to how to hde nformaton based on deep learnng, whch makes t safer and more robust. Tang et al. [Tang, Tan, L et al. (2017)] use generatve adversaral network (GAN) to acheve end-to-end nformaton hdng. Baluja [Baluja (2017)] use neural network to determne the embeddng secret

3 A Fuson Steganographc Algorthm Based on Faster R-CNN 3 nformaton n the locaton of the mage, tran an encoder to embed nformaton. Uchda et al. [Uchda, Naga, Sakazawa et al. (2017)] embed watermark nto the depth neural network model. Therefore, the combnaton of deep learnng and nformaton hdng has become the focus of ths paper. The exstng methods usually adopt a hdden mode to hde the secret nformaton or watermarkng n the entre mage [Xa, Wang, Zhang et al. (2016); Wang, Lan and Sh (2017); Chen, Zhou, Jeon et al. (2017)]. In order to enhance the securty and complexty of nformaton, t s straghtforward to desgn a fuson hdng strategy that employs dfferent steganography algorthms to hde nformaton on dfferent areas. In addton, the complexty of the mage s closely related to human vsual effects. The more complex the mage, the more nformaton t carres, but the vsual nformaton of people does not ncrease wth the ncrease of mage complexty. Usng the effect of human vsual redundancy, the nformaton s hdden n the more complcated area of the mage texture, so as to ncrease the robustness and ant-detectablty of hdng process. Therefore, ths work proposes an approach to detect the more complex area to hde the nformaton by the method of object detecton. We adopt the fuson method of multple steganographc algorthms as multple steganography Algorthms based on ROI (MSA_ROI) to hde the nformaton. 2 Related works 2.1 Object detecton The am of object detecton s to fnd the locaton of all the targets and specfy each target category on a gven mage or vdeo. It s manly dvded nto two tasks, target postonng and target category detecton. In tradtonal methods, the object detecton s manly use sldng wndow framework. The common algorthm s DPM (Deformable Part Model) [Felzenszwalb, Grshck, Mcallester et al. (2010)]. In face detecton, pedestran detecton and other tasks, DPM have acheved good results. But DPM s relatvely complex that the detecton speed s relatvely slow. Wth the development of deep learnng, object detecton has entered a new era. Object detecton methods related to deep learnng can be dvded nto two categores, one s based on the regonal nomnaton, such as R-CNN (Regonbased Convolutonal Neural Networks) [Grshck, Donahue, Darrell et al. (2014); He, Zhang, Ren et al. (2015)], SPP-net [He, Zhang, Ren et al. (2015)], Fast R-CNN [Grshck (2015)], Faster R-CNN, etc. The other s end-to-end approach, Such as YOLO [Redmon, Dvvala, Grshck et al. (2016)] and SSD. At present, Faster R-CNN model s a typcal object detecton model based on deep learnng. From R-CNN, Fast R-CNN to Faster R- CNN whch used n ths paper, the four basc steps of object detecton (regon proposal, feature extracton, classfcaton and rectangles refne regresson) are fnally unfed nto a deep network framework to acheve end-to-end object detecton. 2.2 Spatal doman steganography The current steganography methods are manly dvded nto two types: spatal doman and transform doman. In transform doman, man applcaton s JPEG. JPEG s a type of mage compresson method that dvdes an mage nto several matrces, performs a dscrete cosne varaton on each matrx and transforms the specfc pxel values on the mage completely

4 4 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 nto the frequency doman. Snce human eye s senstve to low frequency but nsenstve to hgh frequency, all the hgh frequency nformaton s elmnated so as to acheve the purpose of mage compresson. Hdden nformaton wll be embedded n the ntermedate frequency area (hgh-frequency area of ant-attack, low-frequency changes n the regon s easy to be perceved by the human eye), whch wll be embedded nformaton dspersed throughout the mage. In spatal doman, embeddng nformaton s changng pxel values drectly. Adaptve steganography s to automatcally select the carrer mage whch s not easly found by the attacker based on the content of the cover mage features on the regon of nterest and embed secret nformaton. Modfyng pxels of the mage causes less dstorton to the mage on the complexty of the rch texture area and the edge area. At the same tme, t s dffcult for the attacker to detect secret nformaton. The steganography that based on the prncple of mnmzng the embedded dstorton cannot only ensure the mnmum dstorton rate but also acheves secret communcaton. Exstng steganography methods that based on the prncple of mnmzng embedded dstorton nclude: HUGO, WOW, S-UNIWARD, MVGG [L, Wang, Huang et al. (2015)] and so on. The ultmate goal s the same: mnmze the dstorton functon and embed t n nosy or complex textures regon of the cover mage S-UNIWARD algorthm S-UNIWARD s a content adaptve steganography based on wavelet transform, whch proposes a general dstorton functon ndependent of the embedded doman. The embeddng dstorton functon of S-UNIWARD as a whole s: D( X, Y) ( ) ( ) 3 n1 n k k 2 Wuv ( X ) Wuv ( Y) (1) W ( X) ( k ) k 1 u1 v1 uv (1) (2) (3) Here, K, K, K represent horzontal, vertcal, dagonal drecton of the flter n three (1) (2) (3) drectons, calculated by K h g T, K g h T, K g g T, where h represents a onedmensonal wavelet decomposton low (hgh) pass flter. Parameter X represents the carrer mage, the mage sze s n 1 n 2, and parameter Y represents the mage after ( k) ( k) embeddng the message. Parameter Wuv ( X ), Wuv ( Y) represent the mage of the carrer mage and the encrypted mage after wavelet transform. Formula (2) s used to calculate the wavelet coeffcents of pxels n three drectons of the orgnal mage. W K * X,(1 u n,1 v n,1 k 3) (2) ( k) ( k) uv uv 1 2 When the wavelet coeffcents of pxels (such as texture regons) wth complex content areas are changed, the dstorton calculated by formula (1) wll be small, ndcatng that the regon s sutable for hdng nformaton. However, when the pxel wavelet coeffcents of the texture smoothng regon are changed, the dstorton wll be very large, ndcatng that secret nformaton should be avoded when embeddng these pxels HUGO algorthm HUGO s consdered to be one of the most secure steganographc technques. It defnes a dstorton functon doman by assgnng costs to pxels based on the effect of embeddng some nformaton wthn a pxel, the space of pxels s condensed nto a feature space usng

5 A Fuson Steganographc Algorthm Based on Faster R-CNN 5 a weghted norm functon. HUGO algorthm s the use of SPAM steganalyss feature desgn dstorton cost functon. It s consdered to be one of the most secure steganographc technques. Accordng to the addtve dstorton functon: n D( X, Y ) x y (3) 1 Here, the constant 0 s a fxed parameter that represents the amount of dstorton that results from a pxel change. When, the pxel s the so-called wet pxel, and the wet pxel does not allow modfcaton durng embeddng. The mnmum expected dstorton functon s: D ( m, n, p) p (4) mn n 1 e Where p s the probablty that the -th pxel changes. Parameter 1 e n ( ) 1,(0 ) s the set of addtve dstorton metrcs formula (3), where {1,..., n}. Parameter m(0 m n) s the number of bts to be passed when usng bnary embeddng operatons WOW algorthm WOW (weght acquston wavelet) s another method of steganography, whch dependng on the complexty of the regon. It wll be covered embeddng nformaton nto an mage. If one area of the mage s more complex than the other, the pxel values n that area wll be modfed. WOW steganography algorthm manly from the perspectve desgn of the dstorton functon. The addtve dstorton functon s: n1 n2 D( X, Y) ( X, Y ) X Y 1 j1 j j j j Y. X Where j are the costs of changng pxel j to j 2.3 Qualty assessment The expermental results n ths paper used the followng ndexes to evaluate the qualty of stego mages Mean Square Error (MSE) The followng expresson s the Mean Square Error (MSE) between the mages AI and BI. Supposng that the pxel value of the bearer mage s AI (, j),0 j N 1 and the pxel value of the correspondng stego mage s BI (, j),0 M 1,0 j N 1, the error mage s e(, j) AI (, j) BI (, j),0 M 1,0 j N 1, then the mean square error s expressed as: (5)

6 6 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 M 1 N 1 M 1 N 1 MSE 1 e j 1 MN MN AI j BI j 2 2, (, ) (, ) (6) 0 j 0 0 j 0 Lower s the MSE, hgher s the smlarty between the mages Peak Sgnal to Nose Rato (PSNR) The followng expresson s the Peak Sgnal to Nose Rato (PSNR) between the mages k AI and BI. Settng AImax 2 1, where K represents a number of bts for all pxels, the PSNR s defned as: 2 2 AI max AImax MN PSNR 101g 101g M 1N 1 db MSE 2 AI(, j) BI(, j) 0 j0 In many vdeo sequences and commercal mage acquston applcatons, often take k = 8. So, for an 8-bt bnary mage, AImax 255, substtutng formula (7) nto: (7) MN PSNR 101g 101 g ( db) M1N1 MSE 2 AI(, j) BI(, j) 0 j0 Hgher s the PSNR, hgher s the smlarty between the mages. (8) Structural Smlarty Index Measure (SSIM) The Structure Smlarty Index Measure (SSIM) between the mages OI and WI whch are of sze M N s gven by the followng expresson. Gven two mages AI and BI, the structural smlarty of the two mages can be found n accordance wth the followng formula: (2 c )(2 c ) SSIM ( AI, BI ) ( )( ) AI BI 1 AIBI AI BI c1 AI BI c2 2 2 Here, AI a s the average of AI, BI s the average of BI, s the varance of AI, AI s BI 2 2 the varance of BI, and AIBI s the covarance of AI and BI. c1 ( k1l), c2 ( k2l) are constants used to mantan stablty. L s the dynamc range of pxel values. k1 0.01, k Structural smlartes range from -1 to 1. When two mages are dentcal, the value of SSIM equals one. 3 Proposed method usng multple algorthms based on Faster R-CNN Ths secton dscusses our steganographc scheme, the models we use and the nformaton each party wshes to conceal or reveal. After layng ths theoretcal groundwork, we present experments supportng our clams. The overall framework for ths artcle s shown n Fg. 1. Frstly, the whole mage s nput nto the CNN model, and then the feature s extracted. Then, the Proposal s generated through the RPN network, the Proposal s mapped to the last layer of convoluton of CNN, and each proposal s made nto a fxed-sze feature maps (9)

7 A Fuson Steganographc Algorthm Based on Faster R-CNN 7 through Rol poolng layer. Fnally, use Softmax classfcaton and boundng box regresson to get the part of the carrer we need. Next, the steganographc algorthms are matched for each of the obtaned carrer parts to fnally obtan stego mages. 3.1 Extract object by Faster R-CNN In the feld of computer vson, object detecton manly solves two problems: the locaton of multple objects on the mage and the categores of each object. Faster R-CNN ntroduced the regon proposal network (RPN) based on Fast R-CNN, replacng the slow search selectve search algorthm. Regon proposal uses nformaton such as the texture, edge, and color n the mage to fnd out the poston where the target on the way may appear beforehand, and can guarantee a hgher recall rate wth fewer wndows selected (a few hundred or even a few thousand). Ths greatly reduces the tme complexty of follow-up operatons, and obtans the canddate wndow than the sldng wndow of hgher qualty. In a sense, Faster R-CNN = RPN + Fast R-CNN. Takng nto account the target detecton s based on the mage texture, edge and determne the target. Ths paper argues that objects selected from the Faster R-CNN are more conducve to hdng nformaton than the background. Therefore, Faster R-CNN s used n the method proposed here. Faster R- CNN's network model s shown n Fg. 2. Faster R-CNN s manly dvded nto four contents: Extract object by Faster R-CNN (a) Proposals... Feature Maps Proposals... Input Image Feature Extracton Regon Proposal Network Rol poolng Classfer Boundng box of Steganographc Algorthm (b) Regons of Objects Matchng Algorthm Pretreatment Fgure 1: Proposed Steganographc archtecture. (a) Target detecton structure based on Faster R-CNN. (b) Steganography algorthm structure for the local area matchng Conv layers: Includng the 13 conv layers +13 relu layers +4 poolng layers, used to extract mage features maps. The feature maps are shared for subsequent RPN layers and full connectvty layers. Regon Proposal Networks: used to generate regon proposals. The RPN network frst passes a 33 convolutonal layer, generatng foreground anchors and boundng box

8 8 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 regresson offsets, respectvely, and then calculates proposals. Anchors belong to foreground or background va softmax functon. Ro Poolng: Ths layer accordng to feature maps and proposals to extract proposal feature maps, nto the subsequent full connecton layer. Classfcaton: Use proposal feature maps to calculate the type of proposal, and then use boundng box regresson to get the fnal exact poston of the test box. Conv layers(13), Relu layers(13), Poolng layers(4) Input Image... Feature Map ROIPoolng Proposal Reshape Softmax Reshape 1x1 3x3 1x1 Bbox_pred Softmax cls_prob Fgure 2: Network model of Faster R-CNN, ncludng convolutonal layers, regon proposal network, RoI poolng and classfcaton The loss functon of Faster R-CNN s: 1 1 L p t L p p p L t t (10) * * * ({ },{ }) cls (, ) reg (, ) Ncls Nreg p s the predcted probablty of anchor beng an object. The ground-truth label: p * 0 1 t t, t, t, t x y w h negatvelable postve lable * box of the predcton. t (11) s a vector representng the 4 parameterzed coordnates of the boundng s the coordnate vector of the ground truth boundng box correspondng to the postve anchor. * cls, * * * Lcls p, p log[ p p (1 p )(1 p )] * Lreg t, t s the regresson loss, calculated usng L t t s the logarthmc loss target and non-target: (12) L t t R t t, where R s the * * reg, ( ) * * smooth L1 functon. pl means that only the foreground anchor( p 1) has a regresson reg * loss, and n other cases there s no ( p 0). The outputs of cls and reg are composed of p andu, respectvely, then normalzed by Ncls and N reg.

9 A Fuson Steganographc Algorthm Based on Faster R-CNN Matchng process of steganographc algorthm From the frst part, we get multple texture complex regons. Next, we need to hde the nformaton by matchng the regons wth dfferent steganographc algorthms as shown n Fg. 3. Frst of all, the proposed method needs to judge whether there s overlap n the target area. If there are overlappng parts, the overlap part should be processed. Then the mage should be grayscale. In the order stage, the hash algorthm s used to sort the target area n each cover mage. Fnally, the sorted target area and the steganography algorthm s sequentally matched to complete the concealment of the secret nformaton. Selecton S-UNIWARD Regons of Objects Grayscale... WOW Stego HUGO Judgement Pretreament Order Matchng Fgure 3: Flow chart of steganographc algorthm matchng process Selecton of overlapped box Frst of all, many of the target frames are overlapped after a target s detected by Faster R- CNN. Secondly, only one steganographc method can be used to hde the nformaton n each target frame. Therefore, t needs to preprocess the overlapped box. As shown n Fg. 4, we take the prncple of maxmum area, for overlappng target box, frst calculate the area of overlapped box, wth a large target box shall preval, small target box by removng the remanng large area as a carrer. (a) (b) (c) (d) Fgure 4: Preprocess for overlappng target box. (a) Target area obtaned by Faster R-CNN. From the graph, we can see that there are overlappng areas n the red border. If the correspondng steganography algorthm s based on the box, t wll cause the overlappng area to hde the nformaton repeatedly. (b) Removng the overlapped area to get the green area. Due to the prncple of choosng maxmum area. (c), the result s shown n the purple area on the choce of maxmum area. The blue and green areas n (d) are the areas of the fnal choce Order of probablty scores

10 10 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 After obtanng the grayscale regonal target maps, we apply dfferent steganographc algorthms to dfferent regons for embeddng nformaton. Each target graph obtaned by Faster R-CNN has a probablty value of Softmax functon. We sort dfferent regons based on ths value. As shown n Fg. 5, there are three boxes n the fgure, from left to rght named box 1, box 2, box 3. The probablty scores of each box are 0.995, 0.968, and Accordng to descendng order, the probablty scores are 0.995, 0.994, and The sorted areas correspond to the boxes 1, 3, and 2, respectvely numbered as 1, 2, and 3. Fgure 5: Target areas of the cover mage. For dfferent regons have dfferent probablty values, from left to rght are 0.995, , Match steganographc algorthms After gettng the order of probablty scores, we use hash algorthm to match steganographc algorthms. The method s usng dvson hash algorthm on the number, takng the remander of 3. When the remander s 0, S-UNIWARD algorthm s used. When the remander s 1, WOW algorthm s used and when the remander s 2, we use the HUGO algorthm. As shown n Fg. 6, the stego mage s used by dfferent matchng steganographc algorthms to embed the nformaton n dfferent regons. Fgure 6: The stego mage 4 Experments All the experments GPU envronment s NVIDIA GTX1080. The experment tranng data set s COCO2014 dataset [Ln, Mare, Belonge et al. (2014)], and the target n the mage s calbrated by exact segmentaton. The mage ncludes 91 categores of targets, 328,000 mages and 2,500,000 labels. All experments used the depth learnng framework Caffe [Ja, Shelhamer, Jeff et al. (2014)]. The network used by the mage feature extracton secton when tranng Faster R-CNN s VGG16 [Smonyan and Zsserman (2014)] network. 4.1 Object extracton

11 A Fuson Steganographc Algorthm Based on Faster R-CNN 11 When we extract the target area, we use the Faster R-CNN model on COCO2014 data set. When tranng, the learnng rate (base_lr) s set to 0.001, gamma s set to 0.1, momentum s set to 0.9 and weght decay s set to After teratons, the network model fle s obtaned. When testng, we use ths network model fle to get the object detecton area box of the mage. Fg. 7 s the mage obtaned after tranng the test mage through the Faster R-CNN model. (a) (b) (c) (d) (e) Fgure 7: The frst row s the fve mages from COCO2014 dataset used n testng process. The second row s the target area after the selecton by Faster R-CNN 4.2 Steganographc process After obtanng the coordnates of the object detecton regon of the mage, three dfferent steganographc algorthms are used to embed the secret nformaton n the regon to obtan the stego mage. The three spatal steganographc algorthms are respectvely S- UNIWARD, HUGO and WOW. Accordng to the dvson hash, dfferent regons correspond to dfferent steganography algorthms. Fg. 8 shows the test results after steganography. Cover Resdual map Stego Fgure 8: The frst row shows the test mages, the second row represents the resdual mages steganography by the proposed method, and the thrd row shows the stego mages

12 12 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 (a) (b) (c) Fgure 9: (a), (b) and (c) are lne charts of the PSNR values, MSE values and SSIM values of the proposed method and the three tradtonal spatal algorthms, respectvely 4.3 Analyss the qualty of stego In ths secton, three dfferent mage qualty ndexes are used to evaluate the qualty of the stego mages, whch are MSE, PSRN, and SSIM. Through experments, t s found that the proposed method s superor to HUGO, S-UNIWARD and WOW steganography n three ndexes of PSNR, MSE and SSIM, ndcatng the least dstorton of the method. As shown n Tab. 1, the PSNR value of SMSA_ROI s hgher than that of HUGO, S-UNIWARD and WOW. The hghest value s db, whch s hgher than 9.24 db n WOW algorthms, ndcatng that the dstorton of MSA_ROI method s the smallest. The maxmum MSE value of the proposed method s , whch s lower than the other three methods, and the mnmum can be as low as The last evaluaton ndex SSIM, the four methods of SSIM value s not much dfference, but MSA_ROI s stll better than the other three algorthms. As shown n Fg. 9, the PSNR value of the proposed method s obvously hgher than the three tradtonal spatal algorthms. The MSE value of the proposed method s obvously lower than the other three methods, and the lowest s The proposed method s hgher than the three tradtonal spatal algorthms on the SSIM value, and the maxmum s up to

13 A Fuson Steganographc Algorthm Based on Faster R-CNN 13 Table 1: Comparson of the proposed algorthm and the sngle spatal algorthm n mage qualty PSNR MSE MSA_ROI_A HUGO_A S-UNIWARD _A WOW_A MSA_ROI_B HUGO_B S-UNIWARD _B WOW_B MSA_ROI_C HUGO_C S-UNIWARD _C WOW_C MSA_ROI_HUGO_D MSA_ROI_ S-UNIWARD _D MSA_ROI_WOW_D HUGO_D S-UNIWARD _D WOW_D MSA_ROI_E HUGO_E S-UNIWARD _E WOW_E Concluson and future work Ths paper has two man contrbutons. The frst one s to combne the object detecton method to select a complex texture regon, whch s sutable for hdng nformaton. The second one ntegrates the exstng multple spatal steganography algorthms nto a cover mage. Experments show that the proposed method s superor to the tradtonal spatal steganography algorthm. Future works to further move ths research ncludes the followng aspects. 1. Hde secret message n the foreground completely. 2. Swtch to dfferent object detecton methods. 3. Adjust the steganography algorthm adaptvely. Acknowledgement: Ths work s supported, n part, by the Natonal Natural Scence Foundaton of Chna under grant numbers U , U , , , , ; n part, by the Jangsu Basc Research Programs-Natural Scence Foundaton under grant numbers BK and BK ; n part, by the Prorty

14 14 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 Academc Program Development of Jangsu Hgher Educaton Insttutons (PAPD) fund; n part, by the Collaboratve Innovaton Center of Atmospherc Envronment and Equpment Technology (CICAEET) fund, Chna. References Baluja, S. (2017): Hdng mages n plan sght: Deep steganography. Advances n Neural Informaton Processng Systems, pp Cao, Y.; Zhou, Z.; Sun, X.; Gao, C. (2018): Coverless nformaton hdng based on the molecular structure mages of materal. Computers, Materals & Contnua, vol. 54, no. 2, pp Chen, B.; Zhou, C.; Jeon, B.; Zheng, Y.; Wang, J. (2017): Quaternon dscrete fractonal random transform for color mage adaptve watermarkng. Multmeda Tools and Applcaton. Da, J.; L, Y.; He, K.; Sun, J. (2016): R-FCN: Object detecton va regon-based fully convolutonal networks. Advances n Neural Informaton Processng Systems, pp Dumtrescu, S.; Wu, X.; Wang, Z. (2003): Detecton of LSB steganography va sample par analyss. IEEE Transactons on Sgnal Processng, vol. 51, no. 7, pp Felzenszwalb, P. F.; Grshck, R. B.; Mcallester, D.; Ramanan, D. (2014): Object detecton wth dscrmnatvely traned part-based models. IEEE Transactons on Pattern Analyss & Machne Intellgence, vol. 47, no. 2, pp Frdrch, J.; Goljan, M. (2012): Practcal steganalyss of dgtal mages: state of the art. Securty and Watermarkng of Multmeda Contents IV, vol. 4675, no. 1, pp Frdrch, J.; Goljan, M.; Du, R. (2001): Steganalyss based on JPEG compatblty. Proc Spe, vol. 4518, pp Grshck, R.; Donahue, J.; Darrell, T.; Malk, J. (2014): Rch feature herarches for accurate object detecton and semantc segmentaton. IEEE Conference on Computer Vson and Pattern Recognton, pp Grshck, R. (2015): Fast R-CNN. IEEE Internatonal Conference on Computer Vson, pp Gurusamy, R.; Subramanam, V. (2017): A machne learnng approach for MRI bran tumor classfcaton. Computers, Materals & Contnua, vol. 53, no. 2, pp He, K.; Zhang, X.; Ren, S.; Sun, J. (2015): Spatal pyramd poolng n deep convolutonal networks for vsual recognton. IEEE Transactons on Pattern Analyss & Machne Intellgence, vol. 37, no. 9, pp Holub, V.; Frdrch, J. (2012): Desgnng steganographc dstorton usng drectonal flters. IEEE Internatonal Workshop on Informaton Forenscs and Securty, vol. 2, no. 4, pp Holub, V.; Frdrch, J.; Denemark, T. (2014): Unversal dstorton functon for steganography n an arbtrary doman. Eurasp Journal on Informaton Securty, vol. 2014, no. 1, pp. 1.

15 A Fuson Steganographc Algorthm Based on Faster R-CNN 15 Ja, Y.; Shelhamer, E.; Donahue, J.; Karayev, S.; Long, J. et al. (2014): Caffe: Convolutonal archtecture for fast feature embeddng. ACM Internatonal Conference on Multmeda, pp Krzhevsky, A.; Sutskever, I.; Hnton, G. E. (2012): Imagenet classfcaton wth deep convolutonal neural networks. Advances n Neural Informaton Processng Systems, vol. 60, no. 2, pp L, B.; Wang, M.; Huang, J.; L, X. (2015): A new cost functon for spatal mage steganography. IEEE Internatonal Conference on Image Processng, pp Ln, T. Y.; Mare, M.; Belonge, S.; Hays, J.; Perona, P. et al. (2014): Mcrosoft COCO: Common objects n context. European Conference on Computer Vson, vol. 8693, pp Lu, W.; Anguelov, D.; Erhan, D.; Szegedy, C.; Reed, S. et al. (2016): SSD: Sngle shot multbox detector. European Conference on Computer Vson, pp Pevný, T.; Fller, T.; Bas, P. (2010): Usng hgh-dmensonal mage models to perform hghly undetectable steganography. Lecture Notes n Computer Scence, vol. 6387, pp Redmon, J.; Dvvala, S.; Grshck, R.; Farhad, A. (2016): You only look once: unfed, real-tme object detecton. IEEE Conference on Computer Vson and Pattern Recognton, IEEE Computer Socety, pp Ren, S.; Grshck, R.; Grshck, R.; Sun, J. (2017): Faster R-CNN: Towards real-tme object detecton wth regon proposal networks. IEEE Transactons on Pattern Analyss & Machne Intellgence, vol. 39, no. 6, pp Smonyan, K.; Zsserman, A. (2014): Very deep convolutonal networks for large-scale mage recognton. Computer Scence. Szegedy, C.; Lu, W.; Ja, Y.; Sermanet, P.; Reed, S. et al. (2015): Gong deeper wth convolutons. IEEE Conference on Computer Vson and Pattern Recognton, pp Tang, W.; Tan, S.; L, B.; Huang, J. (2017): Automatc steganographc dstorton learnng usng a generatve adversaral network. IEEE Sgnal Processng Letters, vol. 24, no. 10, pp Uchda, Y.; Naga, Y.; Sakazawa, S.; Satoh, S. I. (2017): Embeddng watermarks nto deep neural networks. ACM on Internatonal Conference on Multmeda Retreval, pp Wang, J.; Lan, S.; Sh, Y. (2017): Hybrd multplcatve mult-watermarkng n DWT doman. Multdmensonal Systems and Sgnal Processng, vol. 28, no. 2, pp Xa, Z.; Wang, X. H.; Zhang, L. G.; Qn, Z.; Sun, X. et al. (2016): A prvacy-preservng and copy-deterrence content-based mage retreval scheme n cloud computng. IEEE Transactons on Informaton Forenscs and Securty, vol. 11, no. 11, pp Ye, J.; N, J.; Y, Y. (2017): Deep learnng herarchcal representatons for mage steganalyss. IEEE Transactons on Informaton Forenscs & Securty, vol. 12, no. 11, pp Yuan, C.; L, X.; Wu, Q.; L, J.; Sun, X. (2017): Fngerprnt lveness detecton from dfferent fngerprnt materals usng convolutonal neural network and prncpal component analyss. Computers, Materals & Contnua, vol. 53, no. 4, pp

16 16 Copyrght 2018 Tech Scence Press CMC, vol.55, no.1, pp.1-16, 2018 Zhou, Z.; Mu, Y.; Wu, Q. (2018): Coverless mage steganography usng partal-duplcate mage retreval. Soft Computng. Zhou, Z.; Wang, Y.; Wu, Q.; Yang, C.; Sun, X. (2017): Effectve and effcent global context verfcaton for mage copy detecton. IEEE Transactons on Informaton Forenscs and Securty, vol. 12, no. 1, pp Zhou, Z.; Wu, Q.; Huang, F.; Sun, X. (2017): Fast and accurate near-duplcate mage elmnaton for vsual sensor networks. Internatonal Journal of Dstrbuted Sensor Networks, vol. 13, no. 2, pp Zhou, Z.; Wu, Q.; Yang, C.; Sun, X.; Pan, Z. (2017): Coverless mage steganography based on hstograms of orented gradents-based hashng algorthm. Journal of Internet Technology, vol. 18, no. 5, pp Zhou, Z.; Yang, C.; Chen, B.; Sun, X.; Lu, Q. et al. (2016): Effectve and effcent mage copy detecton wth resstance to arbtrary rotaton. IEICE Transactons on Informaton and Systems, vol. E99-D, no. 6, pp

Hybrid Non-Blind Color Image Watermarking

Hybrid Non-Blind Color Image Watermarking Hybrd Non-Blnd Color Image Watermarkng Ms C.N.Sujatha 1, Dr. P. Satyanarayana 2 1 Assocate Professor, Dept. of ECE, SNIST, Yamnampet, Ghatkesar Hyderabad-501301, Telangana 2 Professor, Dept. of ECE, AITS,

More information

An Image Fusion Approach Based on Segmentation Region

An Image Fusion Approach Based on Segmentation Region Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Research and Application of Fingerprint Recognition Based on MATLAB

Research and Application of Fingerprint Recognition Based on MATLAB Send Orders for Reprnts to reprnts@benthamscence.ae The Open Automaton and Control Systems Journal, 205, 7, 07-07 Open Access Research and Applcaton of Fngerprnt Recognton Based on MATLAB Nng Lu* Department

More information

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents

More information

Shape-adaptive DCT and Its Application in Region-based Image Coding

Shape-adaptive DCT and Its Application in Region-based Image Coding Internatonal Journal of Sgnal Processng, Image Processng and Pattern Recognton, pp.99-108 http://dx.do.org/10.14257/sp.2014.7.1.10 Shape-adaptve DCT and Its Applcaton n Regon-based Image Codng Yamn Zheng,

More information

Article Reversible Dual-Image-Based Hiding Scheme Using Block Folding Technique

Article Reversible Dual-Image-Based Hiding Scheme Using Block Folding Technique Artcle Reversble Dual-Image-Based Hdng Scheme Usng Block Foldng Technque Tzu-Chuen Lu, * and Hu-Shh Leng Department of Informaton Management, Chaoyang Unversty of Technology, Tachung 4349, Tawan Department

More information

Deep learning is a good steganalysis tool when embedding key is reused for different images, even if there is a cover source-mismatch

Deep learning is a good steganalysis tool when embedding key is reused for different images, even if there is a cover source-mismatch Deep learnng s a good steganalyss tool when embeddng key s reused for dfferent mages, even f there s a cover source-msmatch Lonel PIBRE 2,3, Jérôme PASQUET 2,3, Dno IENCO 2,3, Marc CHAUMONT 1,2,3 (1) Unversty

More information

Edge Detection in Noisy Images Using the Support Vector Machines

Edge Detection in Noisy Images Using the Support Vector Machines Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More information

A Secured Method for Image Steganography Based On Pixel Values

A Secured Method for Image Steganography Based On Pixel Values A Secured Method for Image Steganography Based On Pxel Values Tarun Gulat #, Sanskrt Gupta * # Assocate Professor, Electroncs and Communcaton Engneerng Department, MMEC, M.M.U., Mullana, Ambala, Haryana,

More information

High Payload Reversible Data Hiding Scheme Using Difference Segmentation and Histogram Shifting

High Payload Reversible Data Hiding Scheme Using Difference Segmentation and Histogram Shifting JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY, VOL. 11, NO. 1, MARCH 2013 9 Hgh Payload Reversble Data Hdng Scheme Usng Dfference Segmentaton and Hstogram Shftng Yung-Chen Chou and Huang-Chng L Abstract

More information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and

More information

Enhanced Watermarking Technique for Color Images using Visual Cryptography

Enhanced Watermarking Technique for Color Images using Visual Cryptography Informaton Assurance and Securty Letters 1 (2010) 024-028 Enhanced Watermarkng Technque for Color Images usng Vsual Cryptography Enas F. Al rawashdeh 1, Rawan I.Zaghloul 2 1 Balqa Appled Unversty, MIS

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An Improved Image Segmentation Algorithm Based on the Otsu Method 3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School

More information

Distortion Function Designing for JPEG Steganography with Uncompressed Side-image

Distortion Function Designing for JPEG Steganography with Uncompressed Side-image Dstorton Functon Desgnng for JPEG Steganography wth Uncompressed Sde-mage Fangjun Huang School of Informaton Scence and Technology, Sun Yat-Sen Unversty, GD 56, Chna huangfj@mal.sysu.edu.cn Jwu Huang School

More information

An Image Compression Algorithm based on Wavelet Transform and LZW

An Image Compression Algorithm based on Wavelet Transform and LZW An Image Compresson Algorthm based on Wavelet Transform and LZW Png Luo a, Janyong Yu b School of Chongqng Unversty of Posts and Telecommuncatons, Chongqng, 400065, Chna Abstract a cylpng@63.com, b y27769864@sna.cn

More information

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram Shape Representaton Robust to the Sketchng Order Usng Dstance Map and Drecton Hstogram Department of Computer Scence Yonse Unversty Kwon Yun CONTENTS Revew Topc Proposed Method System Overvew Sketch Normalzaton

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive Semi-definite Programming Localization in Wireless Sensor Networks Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer

More information

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research

More information

Local Quaternary Patterns and Feature Local Quaternary Patterns

Local Quaternary Patterns and Feature Local Quaternary Patterns Local Quaternary Patterns and Feature Local Quaternary Patterns Jayu Gu and Chengjun Lu The Department of Computer Scence, New Jersey Insttute of Technology, Newark, NJ 0102, USA Abstract - Ths paper presents

More information

ALEXNET FEATURE EXTRACTION AND MULTI-KERNEL LEARNING FOR OBJECT- ORIENTED CLASSIFICATION

ALEXNET FEATURE EXTRACTION AND MULTI-KERNEL LEARNING FOR OBJECT- ORIENTED CLASSIFICATION ALEXNET FEATURE EXTRACTION AND MULTI-KERNEL LEARNING FOR OBJECT- ORIENTED CLASSIFICATION Lng Dng 1, Hongy L 2, *, Changmao Hu 2, We Zhang 2, Shumn Wang 1 1 Insttute of Earthquake Forecastng, Chna Earthquake

More information

Performance Analysis of Data Hiding in MPEG-4 AAC Audio *

Performance Analysis of Data Hiding in MPEG-4 AAC Audio * TSINGHUA SCIENCE AND TECHNOLOGY ISSNll1007-0214ll07/21llpp55-61 Volume 14, Number 1, February 2009 Performance Analyss of Data Hdng n MPEG-4 AAC Audo * XU Shuzheng ( ) **, ZHANG Peng ( ), WANG Pengjun

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Face Detection with Deep Learning

Face Detection with Deep Learning Face Detecton wth Deep Learnng Yu Shen Yus122@ucsd.edu A13227146 Kuan-We Chen kuc010@ucsd.edu A99045121 Yzhou Hao y3hao@ucsd.edu A98017773 Mn Hsuan Wu mhwu@ucsd.edu A92424998 Abstract The project here

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Audio Content Classification Method Research Based on Two-step Strategy

Audio Content Classification Method Research Based on Two-step Strategy (IJACSA) Internatonal Journal of Advanced Computer Scence and Applcatons, Audo Content Classfcaton Method Research Based on Two-step Strategy Sume Lang Department of Computer Scence and Technology Chongqng

More information

An efficient method to build panoramic image mosaics

An efficient method to build panoramic image mosaics An effcent method to buld panoramc mage mosacs Pattern Recognton Letters vol. 4 003 Dae-Hyun Km Yong-In Yoon Jong-Soo Cho School of Electrcal Engneerng and Computer Scence Kyungpook Natonal Unv. Abstract

More information

Detection of an Object by using Principal Component Analysis

Detection of an Object by using Principal Component Analysis Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

Research of Image Recognition Algorithm Based on Depth Learning

Research of Image Recognition Algorithm Based on Depth Learning 208 4th World Conference on Control, Electroncs and Computer Engneerng (WCCECE 208) Research of Image Recognton Algorthm Based on Depth Learnng Zhang Jan, J Xnhao Zhejang Busness College, Hangzhou, Chna,

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation Based Analysis of FAST TCP using OMNET++ Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months

More information

Image Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline

Image Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline mage Vsualzaton mage Vsualzaton mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and Analyss outlne mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and

More information

Collaboratively Regularized Nearest Points for Set Based Recognition

Collaboratively Regularized Nearest Points for Set Based Recognition Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,

More information

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS P.G. Demdov Yaroslavl State Unversty Anatoly Ntn, Vladmr Khryashchev, Olga Stepanova, Igor Kostern EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS Yaroslavl, 2015 Eye

More information

A Hybrid Semi-Blind Gray Scale Image Watermarking Algorithm Based on DWT-SVD using Human Visual System Model

A Hybrid Semi-Blind Gray Scale Image Watermarking Algorithm Based on DWT-SVD using Human Visual System Model A Hybrd Sem-Blnd Gray Scale Image Watermarkng Algorthm Based on DWT-SVD usng Human Vsual System Model Rajesh Mehta r Scence & Engneerng, USICT Guru Gobnd Sngh Indrarprastha Unversty New Delh, Inda rajesh00ust@gmal.com

More information

Modular PCA Face Recognition Based on Weighted Average

Modular PCA Face Recognition Based on Weighted Average odern Appled Scence odular PCA Face Recognton Based on Weghted Average Chengmao Han (Correspondng author) Department of athematcs, Lny Normal Unversty Lny 76005, Chna E-mal: hanchengmao@163.com Abstract

More information

Audio Event Detection and classification using extended R-FCN Approach. Kaiwu Wang, Liping Yang, Bin Yang

Audio Event Detection and classification using extended R-FCN Approach. Kaiwu Wang, Liping Yang, Bin Yang Audo Event Detecton and classfcaton usng extended R-FCN Approach Kawu Wang, Lpng Yang, Bn Yang Key Laboratory of Optoelectronc Technology and Systems(Chongqng Unversty), Mnstry of Educaton, ChongQng Unversty,

More information

PCA Based Gait Segmentation

PCA Based Gait Segmentation Honggu L, Cupng Sh & Xngguo L PCA Based Gat Segmentaton PCA Based Gat Segmentaton Honggu L, Cupng Sh, and Xngguo L 2 Electronc Department, Physcs College, Yangzhou Unversty, 225002 Yangzhou, Chna 2 Department

More information

Key-Selective Patchwork Method for Audio Watermarking

Key-Selective Patchwork Method for Audio Watermarking Internatonal Journal of Dgtal Content Technology and ts Applcatons Volume 4, Number 4, July 2010 Key-Selectve Patchwork Method for Audo Watermarkng 1 Ch-Man Pun, 2 Jng-Jng Jang 1, Frst and Correspondng

More information

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity ISSN(Onlne): 2320-9801 ISSN (Prnt): 2320-9798 Internatonal Journal of Innovatve Research n Computer and Communcaton Engneerng (An ISO 3297: 2007 Certfed Organzaton) Vol.2, Specal Issue 1, March 2014 Proceedngs

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Robust Watermarking for Text Images Based on Arnold Scrambling and DWT-DCT

Robust Watermarking for Text Images Based on Arnold Scrambling and DWT-DCT Internatonal Conference on Mechatroncs Electronc Industral and Control Engneerng (MEIC 015) Robust Watermarkng for Text Images Based on Arnold Scramblng and DWT-DCT Fan Wu College of Informaton Scence

More information

Scale Selective Extended Local Binary Pattern For Texture Classification

Scale Selective Extended Local Binary Pattern For Texture Classification Scale Selectve Extended Local Bnary Pattern For Texture Classfcaton Yutng Hu, Zhlng Long, and Ghassan AlRegb Multmeda & Sensors Lab (MSL) Georga Insttute of Technology 03/09/017 Outlne Texture Representaton

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

Identify the Attack in Embedded Image with Steganalysis Detection Method by PSNR and RGB Intensity

Identify the Attack in Embedded Image with Steganalysis Detection Method by PSNR and RGB Intensity Internatonal Journal of Computer Systems (ISSN: 394-1065), Volume 03 Issue 07, July, 016 Avalable at http://www.jcsonlne.com/ Identfy the Attack n Embedded Image wth Steganalyss Detecton Method by PSNR

More information

Object-Based Techniques for Image Retrieval

Object-Based Techniques for Image Retrieval 54 Zhang, Gao, & Luo Chapter VII Object-Based Technques for Image Retreval Y. J. Zhang, Tsnghua Unversty, Chna Y. Y. Gao, Tsnghua Unversty, Chna Y. Luo, Tsnghua Unversty, Chna ABSTRACT To overcome the

More information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

A Gradient Difference based Technique for Video Text Detection

A Gradient Difference based Technique for Video Text Detection A Gradent Dfference based Technque for Vdeo Text Detecton Palaahnakote Shvakumara, Trung Quy Phan and Chew Lm Tan School of Computng, Natonal Unversty of Sngapore {shva, phanquyt, tancl }@comp.nus.edu.sg

More information

Available online at Available online at Advanced in Control Engineering and Information Science

Available online at   Available online at   Advanced in Control Engineering and Information Science Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced

More information

Fuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches

Fuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches Proceedngs of the Internatonal Conference on Cognton and Recognton Fuzzy Flterng Algorthms for Image Processng: Performance Evaluaton of Varous Approaches Rajoo Pandey and Umesh Ghanekar Department of

More information

Research Article Improved Encrypted-Signals-Based Reversible Data Hiding Using Code Division Multiplexing and Value Expansion

Research Article Improved Encrypted-Signals-Based Reversible Data Hiding Using Code Division Multiplexing and Value Expansion Securty and Communcaton Networks Volume 2018, Artcle ID 1326235, 9 pages https://do.org/10.1155/2018/1326235 Research Artcle Improved Encrypted-Sgnals-Based Reversble Data Hdng Usng Code Dvson Multplexng

More information

A Gradient Difference based Technique for Video Text Detection

A Gradient Difference based Technique for Video Text Detection 2009 10th Internatonal Conference on Document Analyss and Recognton A Gradent Dfference based Technque for Vdeo Text Detecton Palaahnakote Shvakumara, Trung Quy Phan and Chew Lm Tan School of Computng,

More information

Information Hiding Watermarking Detection Technique by PSNR and RGB Intensity

Information Hiding Watermarking Detection Technique by PSNR and RGB Intensity www..org 3 Informaton Hdng Watermarkng Detecton Technque by PSNR and RGB Intensty 1 Neha Chauhan, Akhlesh A. Waoo, 3 P. S. Patheja 1 Research Scholar, BIST, Bhopal, Inda.,3 Assstant Professor, BIST, Bhopal,

More information

A Robust Webpage Information Hiding Method Based on the Slash of Tag

A Robust Webpage Information Hiding Method Based on the Slash of Tag Advanced Engneerng Forum Onlne: 2012-09-26 ISSN: 2234-991X, Vols. 6-7, pp 361-366 do:10.4028/www.scentfc.net/aef.6-7.361 2012 Trans Tech Publcatons, Swtzerland A Robust Webpage Informaton Hdng Method Based

More information

TECHNICAL POINTS ABOUT ADAPTIVE STEGANOGRAPHY BY ORACLE (ASO) 161, rue Ada, 34095, Montpellier Cedex 05, France

TECHNICAL POINTS ABOUT ADAPTIVE STEGANOGRAPHY BY ORACLE (ASO) 161, rue Ada, 34095, Montpellier Cedex 05, France 20th European Sgnal Processng Conference (EUSIPCO 2012) Bucharest, Romana, August 27-31, 2012 TECHNICAL POINTS ABOUT ADAPTIVE STEGANOGRAPHY BY ORACLE (ASO) Sarra Kouder 2, Marc Chaumont 1,2, Wllam Puech

More information

Suppression for Luminance Difference of Stereo Image-Pair Based on Improved Histogram Equalization

Suppression for Luminance Difference of Stereo Image-Pair Based on Improved Histogram Equalization Suppresson for Lumnance Dfference of Stereo Image-Par Based on Improved Hstogram Equalzaton Zhao Llng,, Zheng Yuhu 3, Sun Quansen, Xa Deshen School of Computer Scence and Technology, NJUST, Nanjng, Chna.School

More information

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1 4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:

More information

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines A Modfed Medan Flter for the Removal of Impulse Nose Based on the Support Vector Machnes H. GOMEZ-MORENO, S. MALDONADO-BASCON, F. LOPEZ-FERRERAS, M. UTRILLA- MANSO AND P. GIL-JIMENEZ Departamento de Teoría

More information

A New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1

A New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1 A New Feature of Unformty of Image Texture Drectons Concdng wth the Human Eyes Percepton Xng-Jan He, De-Shuang Huang, Yue Zhang, Tat-Mng Lo 2, and Mchael R. Lyu 3 Intellgent Computng Lab, Insttute of Intellgent

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

Semi-Fragile Watermarking Scheme for Authentication of JPEG Images

Semi-Fragile Watermarking Scheme for Authentication of JPEG Images Tamkang Journal of Scence and Engneerng, Vol. 10, No 1, pp. 5766 (2007) 57 Sem-Fragle Watermarkng Scheme for Authentcaton of JPEG Images Chh-Hung n 1 *, Tung-Shh Su 2 and Wen-Shyong Hseh 2,3 1 Department

More information

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Using Fuzzy Logic to Enhance the Large Size Remote Sensing Images

Using Fuzzy Logic to Enhance the Large Size Remote Sensing Images Internatonal Journal of Informaton and Electroncs Engneerng Vol. 5 No. 6 November 015 Usng Fuzzy Logc to Enhance the Large Sze Remote Sensng Images Trung Nguyen Tu Huy Ngo Hoang and Thoa Vu Van Abstract

More information

CS 534: Computer Vision Model Fitting

CS 534: Computer Vision Model Fitting CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust

More information

AUTOMATIC RECOGNITION OF TRAFFIC SIGNS IN NATURAL SCENE IMAGE BASED ON CENTRAL PROJECTION TRANSFORMATION

AUTOMATIC RECOGNITION OF TRAFFIC SIGNS IN NATURAL SCENE IMAGE BASED ON CENTRAL PROJECTION TRANSFORMATION AUTOMATIC RECOGNITION OF TRAFFIC SIGNS IN NATURAL SCENE IMAGE BASED ON CENTRAL PROJECTION TRANSFORMATION Ka Zhang a, Yehua Sheng a, Pefang Wang b, Ln Luo c, Chun Ye a, Zhjun Gong d a Key Laboratory of

More information

Deep Spatial-Temporal Joint Feature Representation for Video Object Detection

Deep Spatial-Temporal Joint Feature Representation for Video Object Detection sensors Artcle Deep Spatal-Temporal Jont Feature Representaton for Vdeo Object Detecton Baojun Zhao 1,2, Boya Zhao 1,2 ID, Lnbo Tang 1,2, *, Yuq Han 1,2 and Wenzheng Wang 1,2 1 School of Informaton and

More information

Fusion of Deep Features and Weighted VLAD Vectors based on Multiple Features for Image Retrieval

Fusion of Deep Features and Weighted VLAD Vectors based on Multiple Features for Image Retrieval MATEC Web of Conferences, 0500 (07) DTS-07 DO: 005/matecconf/070500 Fuson of Deep Features and Weghted VLAD Vectors based on Multple Features for mage Retreval Yanhong Wang,, Ygang Cen,, Lequan Lang,*,

More information

Classifying Acoustic Transient Signals Using Artificial Intelligence

Classifying Acoustic Transient Signals Using Artificial Intelligence Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)

More information

PERFORMANCE EVALUATION FOR SCENE MATCHING ALGORITHMS BY SVM

PERFORMANCE EVALUATION FOR SCENE MATCHING ALGORITHMS BY SVM PERFORMACE EVALUAIO FOR SCEE MACHIG ALGORIHMS BY SVM Zhaohu Yang a, b, *, Yngyng Chen a, Shaomng Zhang a a he Research Center of Remote Sensng and Geomatc, ongj Unversty, Shangha 200092, Chna - yzhac@63.com

More information

Detecting MP3Stego using Calibrated Side Information Features

Detecting MP3Stego using Calibrated Side Information Features 2628 JOURNAL OF SOFTWARE, VOL. 8, NO. 10, OCTOBER 2013 Detectng P3Stego usng Calbrated Sde Informaton Features Xanmn Yu School of Informaton Scence and Engneerng, Nngbo Unversty Emal: mlhappy1016@163.com

More information

Histogram of Template for Pedestrian Detection

Histogram of Template for Pedestrian Detection PAPER IEICE TRANS. FUNDAMENTALS/COMMUN./ELECTRON./INF. & SYST., VOL. E85-A/B/C/D, No. xx JANUARY 20xx Hstogram of Template for Pedestran Detecton Shaopeng Tang, Non Member, Satosh Goto Fellow Summary In

More information

KEYWORDS: Digital Image Watermarking, Discrete Wavelet Transform, General Regression Neural Network, Human Visual System. 1.

KEYWORDS: Digital Image Watermarking, Discrete Wavelet Transform, General Regression Neural Network, Human Visual System. 1. An Adaptve Dgtal Image Watermarkng Based on Image Features n Dscrete Wavelet Transform Doman and General Regresson Neural Network Ayoub Taher Group of IT Engneerng, Payam Noor Unversty, Broujen, Iran ABSTRACT:

More information

The Study of Remote Sensing Image Classification Based on Support Vector Machine

The Study of Remote Sensing Image Classification Based on Support Vector Machine Sensors & Transducers 03 by IFSA http://www.sensorsportal.com The Study of Remote Sensng Image Classfcaton Based on Support Vector Machne, ZHANG Jan-Hua Key Research Insttute of Yellow Rver Cvlzaton and

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

Research of Multiple Text Watermarks Technique in Electric Power System Texts

Research of Multiple Text Watermarks Technique in Electric Power System Texts Sensors & Transducers 203 by IFSA http://www.sensorsportal.com Research of Multple Text atermarks Technque n Electrc Power System Texts Xao-X XING, Qng CHEN, 2 Lan-X FU School of Optcal-Electrcal and Computer

More information

Fast Feature Value Searching for Face Detection

Fast Feature Value Searching for Face Detection Vol., No. 2 Computer and Informaton Scence Fast Feature Value Searchng for Face Detecton Yunyang Yan Department of Computer Engneerng Huayn Insttute of Technology Hua an 22300, Chna E-mal: areyyyke@63.com

More information

Novel Pattern-based Fingerprint Recognition Technique Using 2D Wavelet Decomposition

Novel Pattern-based Fingerprint Recognition Technique Using 2D Wavelet Decomposition Mathematcal Methods for Informaton Scence and Economcs Novel Pattern-based Fngerprnt Recognton Technque Usng D Wavelet Decomposton TUDOR BARBU Insttute of Computer Scence of the Romanan Academy T. Codrescu,,

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity Journal of Sgnal and Informaton Processng, 013, 4, 114-119 do:10.436/jsp.013.43b00 Publshed Onlne August 013 (http://www.scrp.org/journal/jsp) Corner-Based Image Algnment usng Pyramd Structure wth Gradent

More information

Palmprint Feature Extraction Using 2-D Gabor Filters

Palmprint Feature Extraction Using 2-D Gabor Filters Palmprnt Feature Extracton Usng 2-D Gabor Flters Wa Kn Kong Davd Zhang and Wenxn L Bometrcs Research Centre Department of Computng The Hong Kong Polytechnc Unversty Kowloon Hong Kong Correspondng author:

More information

Face Recognition Based on SVM and 2DPCA

Face Recognition Based on SVM and 2DPCA Vol. 4, o. 3, September, 2011 Face Recognton Based on SVM and 2DPCA Tha Hoang Le, Len Bu Faculty of Informaton Technology, HCMC Unversty of Scence Faculty of Informaton Scences and Engneerng, Unversty

More information

BioTechnology. An Indian Journal FULL PAPER. Trade Science Inc.

BioTechnology. An Indian Journal FULL PAPER. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 0974-74 Volume 0 Issue BoTechnology 04 An Indan Journal FULL PAPER BTAIJ 0() 04 [684-689] Revew on Chna s sports ndustry fnancng market based on market -orented

More information

A high precision collaborative vision measurement of gear chamfering profile

A high precision collaborative vision measurement of gear chamfering profile Internatonal Conference on Advances n Mechancal Engneerng and Industral Informatcs (AMEII 05) A hgh precson collaboratve vson measurement of gear chamferng profle Conglng Zhou, a, Zengpu Xu, b, Chunmng

More information

Lecture 13: High-dimensional Images

Lecture 13: High-dimensional Images Lec : Hgh-dmensonal Images Grayscale Images Lecture : Hgh-dmensonal Images Math 90 Prof. Todd Wttman The Ctadel A grayscale mage s an nteger-valued D matrx. An 8-bt mage takes on values between 0 and 55.

More information

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

Face Recognition University at Buffalo CSE666 Lecture Slides Resources: Face Recognton Unversty at Buffalo CSE666 Lecture Sldes Resources: http://www.face-rec.org/algorthms/ Overvew of face recognton algorthms Correlaton - Pxel based correspondence between two face mages Structural

More information

Research Article High Capacity Reversible Watermarking for Audio by Histogram Shifting and Predicted Error Expansion

Research Article High Capacity Reversible Watermarking for Audio by Histogram Shifting and Predicted Error Expansion e Scentfc World Journal, Artcle ID 656251, 7 pages http://dx.do.org/1.1155/214/656251 Research Artcle Hgh Capacty Reversble Watermarkng for Audo by Hstogram Shftng and Predcted Error Expanson Fe Wang,

More information

Image Matching Algorithm based on Feature-point and DAISY Descriptor

Image Matching Algorithm based on Feature-point and DAISY Descriptor JOURNAL OF MULTIMEDIA, VOL. 9, NO. 6, JUNE 2014 829 Image Matchng Algorthm based on Feature-pont and DAISY Descrptor L L School of Busness, Schuan Agrcultural Unversty, Schuan Dujanyan 611830, Chna Abstract

More information

Fuzzy C-Means Initialized by Fixed Threshold Clustering for Improving Image Retrieval

Fuzzy C-Means Initialized by Fixed Threshold Clustering for Improving Image Retrieval Fuzzy -Means Intalzed by Fxed Threshold lusterng for Improvng Image Retreval NAWARA HANSIRI, SIRIPORN SUPRATID,HOM KIMPAN 3 Faculty of Informaton Technology Rangst Unversty Muang-Ake, Paholyotn Road, Patumtan,

More information

Classification Method in Integrated Information Network Using Vector Image Comparison

Classification Method in Integrated Information Network Using Vector Image Comparison Sensors & Transducers 2014 by IFSA Publshng, S. L. http://www.sensorsportal.com Classfcaton Method n Integrated Informaton Network Usng Vector Image Comparson Zhou Yuan Guangdong Polytechnc Normal Unversty

More information

Face Recognition using 3D Directional Corner Points

Face Recognition using 3D Directional Corner Points 2014 22nd Internatonal Conference on Pattern Recognton Face Recognton usng 3D Drectonal Corner Ponts Xun Yu, Yongsheng Gao School of Engneerng Grffth Unversty Nathan, QLD, Australa xun.yu@grffthun.edu.au,

More information