Modern Object Detection. Most slides from Ali Farhadi

Size: px

Start display at page:

Download "Modern Object Detection. Most slides from Ali Farhadi"

Maria Rose
5 years ago
Views:

1 Modern Object Detection Most slides from Ali Farhadi

2 Comparison of Classifiers assuming x in {0 1} Learning Objective Training Inference Naïve Bayes maximize j i logp + logp ( x y ; θ ) ( y ; θ ) i ij 0 i j θ kj = i δ ( x = 1 y = k) i ij δ ( y = k) + Kr i i + r θ T 1 x + θ whereθ T 0 1 j θ 0 j ( 1 x) > 0 P ( x j = 1 y = 1 = log ), P( x j = 1 y = 0) P( x j = 0 y = 1) = log P( x = 0 y = 0) j Logistic Regression maximize where P ( P( y x, θ) ) log i + λ θ i T ( y x, θ) = 1/ ( 1+ exp( y θ x ) i i Gradient ascent θ T x > t Linear SVM minimize such that 1 λ ξi + θ i 2 T y θ x 1 ξ i i i, ξ 0 i Quadratic programming or subgradient opt. θ T x > t Kernelized SVM complicated to write Quadratic programming ( xˆ, x) yiα ik i > 0 i Nearest Neighbor most similar features same label Record data y i where i = argmin i K ( x ˆ,x) i

3 Image Categorization Training Images Training Image Features Training Labels Classifier Training Trained Classifier Test Image Image Features Testing Trained Classifier Prediction Outdoor

4 Example: Dalal-Triggs pedestrian detector 1. Extract fixed-sized (64x128 pixel) window at each position and scale 2. Compute HOG (histogram of gradient) features within each window 3. Score the window with a linear SVM classifier 4. Perform non-maxima suppression to remove overlapping detections with lower scores Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

5 Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

6 Tested with RGB LAB Grayscale Slightly better performance vs. grayscale

7 Outperforms centered diagonal uncentered cubic-corrected Sobel Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

8 Histogram of gradient orientations Orientation: 9 bins (for unsigned angles) Histograms in 8x8 pixel cells Votes weighted by magnitude Bilinear interpolation between cells Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

9 Normalize with respect to surrounding cells Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

10 # orientations X= # features = 15 x 7 x 9 x 4 = 3780 # cells # normalizations by neighboring cells Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

11 Training set

12 SVM pos w neg w Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

13 pedestrian Slides by Pete Barnum Navneet Dalal and Bill Triggs, Histograms of Oriented Gradients for Human Detection, CVPR05

14 Detection examples

16 Each window is separately classified

18 What about this one? Can the model we trained for pedestrians detect the person in this image?

19 Specifying an object model Statistical Template in Bounding Box Object is some (x,y,w,h) in image Features defined wrt bounding box coordinates Image Template Visualization Images from Felzenszwalb

23 When do statistical templates make sense? Caltech 101 Average Object Images

24 Deformable objects Images from Caltech-256 Slide Credit: Duan Tran

25 Deformable objects Images from D. Ramanan s dataset Slide Credit: Duan Tran

26 Parts-based Models Define objects by collection of parts modeled by 1. Appearance 2. Spatial configuration Slide credit: Rob Fergus

27 Explicit Models Hybrid template/parts model Detections Template Visualization Felzenszwalb et al. 2008

28 How to model spatial relations? Many others... O(N 6 ) O(N 2 ) O(N 3 ) O(N 2 ) Fergus et al. 03 Fei-Fei et al. 03 Leibe et al. 04, 08 Crandall et al. 05 Fergus et al. 05 Crandall et al. 05 Felzenszwalb & Huttenlocher 05 Csurka 04 Vasconcelos 00 Bouchard & Triggs 05 Carneiro & Lowe 06 from [Carneiro & Lowe, ECCV 06]

29 Tree-shaped model

30 Pictorial Structures Model Part = oriented rectangle Spatial model = relative size/orientation Felzenszwalb and Huttenlocher 2005

31 Pictorial Structures Model Appearance likelihood Geometry likelihood

32 Part representation Background subtraction

33 Pictorial Structures

34 Results for person matching 34

35 Results for person matching 35

36 Enhanced pictorial structures BMVC 2009

37 Deformable Latent Parts Model Useful parts discovered during training Detections Template Visualization Felzenszwalb et al. 2008

38 Score = F 0.Φ(p 0,H) + Σ F i.φ(p i,h) - Σ d i.φ d (x,y) 38

State-of-the-art Detector: Deformable Parts Model

part-based models (pictorial structures) 3.

39 State-of-the-art Detector: Deformable Parts Model (DPM) Lifetime Achievement 1. Strong low-level features based on HOG 2. Efficient matching algorithms for deformable part-based models (pictorial structures) 3. Discriminative learning with latent variables (latent SVM) Felzenszwalb 39 et al., 2008, 2010, 2011, 2012

40 40

41 41

42 Car 42

43 43

44 Cat 44

45 45

46 Person riding horse

47 Person riding bicycle

48 48

49 Structure Recognition using Visual Phrases, CVPR 2011

Object Category Detection. Slides mostly from Derek Hoiem

Object Category Detection. Slides mostly from Derek Hoiem Object Category Detection Slides mostly from Derek Hoiem Today s class: Object Category Detection Overview of object category detection Statistical template matching with sliding window Part-based Models