Networked Telepresence System Using Web Browsers and Omni-directional Video Streams

Similar documents
A Networked Telepresence System with an Internet Automobile and its Transmission Issue

Real-time Generation and Presentation of View-dependent Binocular Stereo Images Using a Sequence of Omnidirectional Images

Outdoor Scene Reconstruction from Multiple Image Sequences Captured by a Hand-held Video Camera

AR Cultural Heritage Reconstruction Based on Feature Landmark Database Constructed by Using Omnidirectional Range Sensor

Super Wide Viewing for Tele-operation

Extrinsic Camera Parameter Recovery from Multiple Image Sequences Captured by an Omni-directional Multi-camera System

Localization of Wearable Users Using Invisible Retro-reflective Markers and an IR Camera

3D Reconstruction of Outdoor Environments from Omnidirectional Range and Color Images

DETECTION OF 3D POINTS ON MOVING OBJECTS FROM POINT CLOUD DATA FOR 3D MODELING OF OUTDOOR ENVIRONMENTS

Estimating Camera Position And Posture by Using Feature Landmark Database

Real-Time Geometric Registration Using Feature Landmark Database for Augmented Reality Applications

PanoMOBI: Panoramic Mobile Entertainment System

Omni-directional Multi-baseline Stereo without Similarity Measures

3D Modeling of Outdoor Environments by Integrating Omnidirectional Range and Color Images

Camera Recovery of an Omnidirectional Multi-camera System Using GPS Positions

Remote Reality Demonstration

A System for Visualization and Summarization of Omnidirectional Surveillance Video

NTSC/PAL. Network Camera SNC-P1

Dense 3-D Reconstruction of an Outdoor Scene by Hundreds-baseline Stereo Using a Hand-held Video Camera

Background Estimation for a Single Omnidirectional Image Sequence Captured with a Moving Camera

This letter contains latest information about the above mentioned software version.

A Wearable Augmented Reality System Using an IrDA Device and a Passometer

SRG-300SE/301SE/201SE

Evaluation of Image Processing Algorithms on Vehicle Safety System Based on Free-viewpoint Image Rendering

Product Specification

This letter contains latest information about the above mentioned software version.

Construction of an Immersive Mixed Environment Using an Omnidirectional Stereo Image Sensor

A WALKTHROUGH REMOTE VIEW SYSTEM WITH AN OMNIDIRECTIONAL CAMERA

Computer Vision Technology Applied to MR-Based Pre-visualization in Filmmaking

SDA Application Software

Foscam. Wireless IP Cameras. Foscam Digital Technologies LLC

P AL/NTSC. Network Mini Dome Cameras. SNC-DF Series SNC-DF40N SNC-DF40P SNC-DF70N SNC-DF70P. Sold by: Toll Free: (877)

High performance surveillance. 24/7.

Introduction to LAN/WAN. Application Layer 4

DMAC-VMS-1CH, 4CH, 8CH, 16CH, 24CH, 32CH, 64CH, 128CH

Contents. 1 Introduction... 4

3D MODELING OF OUTDOOR SCENES BY INTEGRATING STOP-AND-GO AND CONTINUOUS SCANNING OF RANGEFINDER

This letter contains latest information about the above mentioned software version.

Changing your point of view SNC-P5. Sony Network Camera.

Megapixel Resolution for Excellent Video Quality. Multi-Streaming: Simultaneous H MPEG4 + 3GPP + Motion-JPEG

Quick Reference Guide

SYSTEM FOR ACTIVE VIDEO OBSERVATION OVER THE INTERNET

CMOS/Mega-Pixel CMOS/CCD Internet Camera ICA-230 / ICA-M230 / ICA-501. Quick Installation Guide

Development of Low-Cost Compact Omnidirectional Vision Sensors and their applications

Control Center 15 Performance Reference Guide

Quick Start Guide. MotionDV STUDIO 5.6. Cover

Real-time Integral Photography Holographic Pyramid using a Game Engine

Day & Night 6MP 360 Panorama Dome IR IP Camera

This letter contains latest information about the above mentioned software version.

DEVELOPMENT OF REAL TIME 3-D MEASUREMENT SYSTEM USING INTENSITY RATIO METHOD

CMS USER S MANUAL VER:

Wireless Day / Night Cloud Camera TV-IP751WIC (v1.0r)

SOHO NETWORK IP CAMERA USER MANUAL

MAXPRO NVR XE (Xpress Edition)

Efficient Acquisition of Human Existence Priors from Motion Trajectories

MAXPRO. NVR Software NETWORK VIDEO RECORDING SOLUTION

SSL Encryption Secure network connections using data encryption between the camera and the network

ONSIP OS101 User s Guide D1BD

Network Camera SNC-CS3N SNC-CS3P

A Streaming Method with Trick Play on Time Division based Multicast

Image and video processing

A 100Hz Real-time Sensing System of Textured Range Images

Network Projector Operation Guide

Real-time Video Surveillance for Large Scenes

Lesson 5: Multimedia on the Web

Day & Night 1080P HD Auto Focus Dome IP Camera

Internet Camera. User Manual

SNC-DF40P High Resolution Minidome Color Camera with 10/100 Base T Ethernet

SunEyes V Series IP Camera P2P CMS user manual

CMS USER S MANUAL VER:

F65RAW Viewer Operation Guide

IP Camera. User Manual

Construction of Feature Landmark Database Using Omnidirectional Videos and GPS Positions

CMS PRO Software Manual - Beta. Copyright Digimerge 2010

Wired / Wireless IR IP Camera ICA-108 / ICA-108W. Quick Installation Guide

A Stereo Vision-based Mixed Reality System with Natural Feature Point Tracking

Area View Kit DC AVK. Features. Applications

A&E SPECIFICATIONS. TeleEye MX Series Network Camera Model MX925-HD

Index Generation and Advanced Search Functions for. Muitimedia Presentation Material. Yahiko Kambayashi 3 Kaoru Katayama 3 Yasuhiro Kamiya 3

This letter contains latest information about the above mentioned software version.

Artec Leo. A smart professional 3D scanner for a next-generation user experience

Megapixel PoE Day / Night PTZ Internet Camera TV-IP672PI (V1.0R)

MAXPRO NVR XE (Xpress Edition)

Mobile Robot Control Interface Using Augmented Free-viewpoint Image Generation

HDMI based Video Conference Device Recording

Implementation of Omni-directional Image Viewer Program for Effective Monitoring

Simultaneous Tele-visualization of Construction Machine and Environment Using Body Mounted Cameras

Chapter 28. Multimedia

2 x Maximum Display Monitor(s) support MHz Core Clock 28 nm Chip 384 x Stream Processors. 145(L)X95(W)X26(H) mm Size. 1.

A Wearable Augmented Reality System Using Positioning Infrastructures and a Pedometer

ipad Support Android Support TeleDRM TM CDN CDN Support Billing system CMS SMS Multi-Screen PRODUCT BROCHURE

Occlusion Detection of Real Objects using Contour Based Stereo Matching

IVC-8371P. 4 Channel Hardware Codec MPEG-4 Video/Audio Capture Card

Local Playback Software User Manual

MR-Mirror: A Complex of Real and Virtual Mirrors

WIFI VS BLUETOOTH Bluetooth. Wi-Fi.

Network Video Recorder: NUUO NVR Titan NT-4040(R)

A Simple Interface for Mobile Robot Equipped with Single Camera using Motion Stereo Vision

Cisco Jabber Video for TelePresence

Cisco WVC210 Wireless-G Pan Tilt Zoom (PTZ) Internet Video Camera: 2-Way Audio Cisco Small Business Video Surveillance Cameras

Transcription:

Networked Telepresence System Using Web Browsers and Omni-directional Video Streams Tomoya Ishikawa a, Kazumasa Yamazawa a, Tomokazu Sato a, Sei Ikeda a, Yutaka Nakamura a, Kazutoshi Fujikawa a, Hideki Sunahara a, Naokazu Yokoya a a Graduate School of Information Science, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara, 630-0192 Japan ABSTRACT In this paper, we describe a new Telepresence system which enables a user to look around a virtualized real world easily in network environments. The proposed system includes omni-directional video viewers on web browsers and allows the user to look around the omni-directional video contents on the web browsers. The omni-directional video viewer is implemented as an Active-X program so that the user can install the viewer automatically only by opening the web site which contains the omni-directional video contents. The system allows many users at different sites to look around the scene just like an interactive TV using a multi-cast protocol without increasing the network traffic. We describe the implemented system and the experiments using live and stored video streams. In the experiment with stored video streams, the system users an omni-directional multi-camera system for video capturing. We can look around high resolution and high quality video contents. We have also implemented a view-dependent presentation with a headmounted display (HMD) and a gyro sensor for realizing more rich presence. In the experiment with live video streams, a car-mounted omni-directional camera acquires omni-directional video streams surrounding the car, running in an outdoor environment. The acquired video streams are transferred to the remote site through the wireless and wired network using multi-cast protocol. We can see the live video contents freely in arbitrary direction. Keywords: Network, Telepresence, Web Browser, Omni-directional Video Stream, Multi-cast 1. INTRODUCTION Recently, telepresence which allows us to experience a remote site through a virtualized real world has been investigated 1. The technology can be widely applied to a number of fields such as entertainment, medical equipment and education. In previous work, we have proposed telepresence systems using omni-directional images which enable a user to see the virtualized dynamic scene freely in arbitrary direction 2-3. The users can look around a wide area in a virtualized remote site without significant delay from the rotation of user s head to the presentation of images. The conventional systems were realized without network and with a stand-alone system that uses an immersive display. Therefore, it was difficult for many users to use the system simultaneously at different places. In this paper, we describe a new telepresence system which enables the user to look around a virtualized real world easily in network environments. The new system includes omni-directional video viewers on web browsers and allows the user to look around the omni-directional video contents on the web browsers. The omni-directional video viewer is implemented as an Active-X program. The user can install the viewer automatically only by opening the web site which contains the omni-directional video contents. Moreover, we describe the implemented system and the experiments using live and stored video streams. In the experiment with stored video streams, the system uses an omni-directional multi-camera system Ladybug for video capturing. We can look around high resolution and high quality video contents. Further author information: T.I, K.Y, T.S, S.I, N.Y: E-mail: {tomoya-i, yamazawa, tomoka-s, sei-i, yokoya}@is.naist.jp Y.N, K.F, H.S: E-mail: {yutaka-n, Fujikawa, suna}@itc.naist.jp

We have also implemented a view-dependent presentation with an HMD and a gyro sensor for realizing more rich presence. In the experiment with live video streams, a car-mounted omni-directional camera acquires omni-directional video streams surrounding the car, running in an outdoor environment. The acquired video streams are transferred to the remote site through the wireless and wired network. Then, we can see the live video contents freely in arbitrary direction. The system allows many users at different site to look around the scene just like an interactive TV using a multi-cast protocol without increasing the network traffic. Section 2 describes the proposed system. Section 3 describes experiments of the system with live video and stored video. Finally, Section 4 summarizes the present work. 2. NETWORKED TELEPRESENCE SYSTEM USING WEB BROWSERS AND OMNI- DIRECTIONAL VIDEO STREAMS The schema of proposed system is illustrated in Fig. 1. Omni-directional videos are stored in remote server and are acquired by the viewer which is started by a web browser. The user looks around the omni-directional video contents on the web browser. 2.1. Omni-directional video viewer using web browser A web browser is one of the most popular network applications. Especially, Internet Explorer installed on Windows machines can execute various application programs by a JAVA applet or Active-X for providing users with interactive contents on a web page. Moreover, the JAVA applet and the Active-X programs can be easily distributed by an automatic install function. Thus we implement an omni-directional video viewer for telepresence on a web browser in this study. Fig. 2 shows the structure of the omni-directional video viewer. The omni-directional video viewer which shows the user omni-directional video contents needs the functions of GPU (graphic processor unit) and is implemented by Active-X. When the user access the web page which provides an omni-directional video contest, the omni-directional video viewer implemented by ActiveX is installed automatically. The omni-directional video viewer is started by the web browser and is provided the parameters such as an omni-directional camera type, parameters of camera, file-path of content, and so on. The parameters are embedded in a HTML file by a content provider and so the user does not need the care of them. Firstly, the omni-directional video viewer accesses the omni-directional video stream or file using DirectShow functions. DirectShow functions can handle the various video formats such as AVI, MPEG, WMV, and so on, and acquire the image of video. Secondly, the acquired image is transferred to the OpenGL function. We use OpenGL functions with GPU functions for converting the omni-directional image to the common perspective image. OpenGL function use a hardware texture mapping of GPU function to a sphere object for converting image in real-time by the method 2. Finally, the omni-directional video viewer draws the converted image to the window of the web browser. These functions are synchronized with the frame rate of inputted video. The user can look around the omni-directional video by using a mouse-drag operation. The omni-directional video viewer is installed only by opening the web page, and the omni-directional video can be seen easily. In the other implementation, the user can look around the omni-directional video through a HMD (Head Mounted Display) with a gyro sensor. When the user uses a HMD, the omni-directional video viewer draws the video on full screen by setting the full screen parameter. 2.2. Omni-directional video contents There are two kinds of omni-directional video contents in present implementation: stored video contents encoded in advance and live video contents encoded in real-time. The stored video contents mainly consist of high-resolution omnidirectional videos heavy for network. Note that stored video contents can be provided as an on-demand-service. The live video contents are used for the purpose of providing multiple users with the same contents simultaneously just like TV

Omni-directional video server Network Omni-directional local file Fig. 1. Overview of proposed system Web browser User Omni-directional video stream or file Web browser HTML file embedded with parameters of viewer Parameters of viewer User Input (view direction) DirectShow Loading of video Acquisition of frame OpenGL Texture mapping to sphere object using GPU function ActiveX Omni-directional image Fig. 2. Structure of omni-directional video viewer broadcasting. It is difficult to transfer a high-resolution video because of the limit of standard network bandwidth. The user can see the live video contents acquired by an omni-directional camera and transferred immediately. It can be transferred to many sites by multi-cast without increasing the network load. 3. EXPERIMENTS We have implemented the proposed system and experiments with both stored video contents and live video contents. Stored video contents consist of high-resolution videos obtained by using an omni-directional multi-camera system. Live video contents are acquired by using an omni-directional camera mounted on a car and that are transferred via wireless and wired network in real time. A) Stored video contents We acquired the omni-directional video by an omni-directional multi-camera system; Ladybug 3 (see Fig.3) and stored it in a PC for presentation (see Table 1). The camera unit of Ladybug consists of six cameras (Fig. 3(left)): Five configured in a horizontal ring and one pointing vertically. Fig. 3(right) shows a storage unit which consists of an array of HDD. The camera system can acquire video covering 75% of full spherical view (see Fig. 4). The acquired video has the size of 3840x1920 pixels and is captured at 15fps. In this experiment, we shrink the video to 1024x512 pixels

Fig. 3. Omni-directional multi-camera system; Ladybug Fig. 4. High-resolution omni-directional image from Ladybug Table 1. PC for presentation of stored video contents CPU Pentium4 2GHz Memory Graphics card OS 512MB ATI RADEON9700pro Windows XP because of the limit of HDD-access-speed of the PC. We used MPEG-1 video and MPEG-1 layer2 sounds for the formats of the video. Fig. 5 shows a widow shot of the web browser. The user can look around the scene on the web browser. The PC can playback the video at 30fps. The user can pause and fast-forward the video with the stored video contents. We have also implemented a view-dependent presentation system (see Fig. 6) with a HMD and a gyro sensor for realizing more rich presence. The gyro sensor is InterTrax2 made by INTERSENSE. It can acquire the user viewdirection at 256Hz. The user can look around the omni-directional scene without significant delay.

Fig. 5. Window shots of omni-directional video viewer with stored video content Gyro sensor HMD Fig. 6. HMD(Head Mounted Display) with gyro sensor B) Live video contents In the experiment of live video contents, the system consists of a car which mounts omni-directional camera HyperOmni Vision 4, multi-cast relay server of video, omni-directional video viewer, and network (see Fig. 7). The car mounting the omni-directional camera acquires omni-directional progressive video surrounding the car, running in our campus. The acquired omni-directional video transferred to a PC for encoding in the through i.link. The PC encodes the omni-directional video (640x480 pixels, 30fps) to Windows Media Format (1Mbps) by Windows Media Encoder 5. The encoded omni-directional video is transferred to the indoor relay server through IEEE802.11a or g network via wireless network. Table 2 describes the configuration of the system for acquiring omni-directional video streams. Fig. 8 shows the car which mounts the omni-directional camera and the system which mounted on the car.

The transferred omni-directional video is received by the multi-cast relay server of omni-directional video. The relay server distributes the omni-directional video by multi-cast such as RTSP protocol. The distributed omni-directional video is seen by using the same omni-directional video viewers as for stored video contents on the web browsers. When many viewers receive the omni-directional video, the load of network does not increase because of using not unicast but multi-cast. In the experiment, actually four PC received the distributed omni-directional video. The four users could look around the scene in arbitrary directions. Fig. 9 and Fig. 10 show examples of windows shots of the omni-directional video viewer. The video is displayed on the web browser at 30fps. The time delay between the acquisition and the presentation omni-directional video is 10 seconds. In this time, the both transmitting and receiving network loads of the relay server are 1Mbps. When the number of received omni-directional video viewer increased, the network load did not increase. Omni-directional video viewer Multicast relay server of omni-directional video Omni-directional video (Multicast) Omni-directional camera Omni-directional video (ilink) PC for wireless PC for video acquisition and encoding Car for omni-directional video acquisition Fig. 7. Telepresence system with omni-directional camera mounted on car

Table 2. Facilities for omni-directional video acquisition in outdoor environments SONY DCR-TRV900 + Omni-directional camera Hyperboloidal mirror (field of view : 30 degrees upper) Pentium4 2.53GHz PC for video acquisition and Memory 1GB encoding Windows XP Wireless network IEEE802.11a and g Car Nissan ELGRAND (see Fig.8) Fig. 8. Appearance of car for omni-directional video acquisition common perspective image Panorama image omni-directional image Fig. 9. Window shots of omni-directional video viewer (user 1)

Fig. 10. Window shots of omni-directional video viewer (other users) 4. CONCLUSION We have developed a new networked telepresence system which easily enables multiple users to look around a remote scene with omni-directional camera. The system uses a web browser, and enables users to see omni-directional video such as common video. In the experiment of stored video contents, the user could see the high-resolution omnidirectional video was distributed through wireless and wired network by multi-cast protocol, and multiple users could look around the scene in arbitrary directions in real time. In the experiment of live video contents, the omni-directional camera was NTSC. The resolution of the omni-directional video was low. Thus the presence was not rich enough. On the other hand, in the experiment of stored video contents, and omni-directional multi-camera system was employed for acquiring a high-resolution video. It was not suitable for distributing an omni-directional live video. Therefore omnidirectional camera should be high resolution and should not be multi-camera system. The omni-directional HD camera can acquire omni-directional high-resolution live video. In the experiment of live video contents, the delay between the acquisition and the presentation omni-directional video is 10 seconds. It is difficult to use the system for communication with a remote user. In future work, we should reduce the delay in transmitting an omni-directional video stream. REFERENCES 1. S.Moezzi, Special issue on immersive telepresence, IEEE Multimedia, vol.4, pp.17-56, 1997. 2. Y.Onoe, K.Yamazawa, H.Takemura, and N.Yokoya, Telepresence by Real-time View-dependent Image Generation from Omnidirectional Video Streams, Computer Vision and Image Understanding, vol.71, pp154-165, 1998. 3. S.Ikeda, T.Sato, M.Kanbara, and N.Yokoya, Immersive Telepresence System Using High-resolution Omnidirectional Movies and a Locomotion Interface, Proc. SPIE Electronic Imaging, vol.5291, 2004. 4. K.Yamazawa, Y.Yagi, and M.Yachida, Omnidirectional imaging with hyperboloidal projection, Proc. Int. Conf. on Intelligent Robots and Systems, vol.2, 1993. 5. Microsoft Corporation, Windows Media Encoder 9 Series, http://www.microsoft.com/windowsmedia/9series/encoder/default.aspx