Real-Time Hand Gesture Recognition with Kinect for Playing Racing Video Games 2014 International Joint Conference on Neural Networks (IJCNN) July 6-11,

Slides:



Advertisements
Similar presentations
Advanced Image Processing Student Seminar: Lipreading Method using color extraction method and eigenspace technique ( Yasuyuki Nakata and Moritoshi Ando.
Advertisements

Automatic Color Gamut Calibration Cristobal Alvarez-Russell Michael Novitzky Phillip Marks.
By: Mani Baghaei Fard.  During recent years number of moving vehicles in roads and highways has been considerably increased.
QR Code Recognition Based On Image Processing
Road-Sign Detection and Recognition Based on Support Vector Machines Saturnino, Sergio et al. Yunjia Man ECG 782 Dr. Brendan.
Robust Part-Based Hand Gesture Recognition Using Kinect Sensor
Wrist Recognition and the Center of the Palm Estimation Based on Depth Camera Zhengwei Yao ; Zhigeng Pan ; Shuchang Xu Virtual Reality and Visualization.
Xin Zhang, Zhichao Ye, Lianwen Jin, Ziyong Feng, and Shaojie Xu
Multi-scenario Gesture Recognition Using Kinect Student : Sin- Jhu YE Student Id : MA Computer Engineering & Computer Science University of Louisville.
A Modified EM Algorithm for Hand Gesture Segmentation in RGB-D Data 2014 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) July 6-11, 2014, Beijing,
Real-Time Accurate Stereo Matching using Modified Two-Pass Aggregation and Winner- Take-All Guided Dynamic Programming Xuefeng Chang, Zhong Zhou, Yingjie.
Generation of Virtual Image from Multiple View Point Image Database Haruki Kawanaka, Nobuaki Sado and Yuji Iwahori Nagoya Institute of Technology, Japan.
Vision Based Control Motion Matt Baker Kevin VanDyke.
A Robust Method of Detecting Hand Gestures Using Depth Sensors Yan Wen, Chuanyan Hu, Guanghui Yu, Changbo Wang Haptic Audio Visual Environments and Games.
Adviser : Ming-Yuan Shieh Student ID : M Student : Chung-Chieh Lien VIDEO OBJECT SEGMENTATION AND ITS SALIENT MOTION DETECTION USING ADAPTIVE BACKGROUND.
A KLT-Based Approach for Occlusion Handling in Human Tracking Chenyuan Zhang, Jiu Xu, Axel Beaugendre and Satoshi Goto 2012 Picture Coding Symposium.
Virtual Dart: An Augmented Reality Game on Mobile Device Supervisor: Professor Michael R. Lyu Prepared by: Lai Chung Sum Siu Ho Tung.
Recognition of Traffic Lights in Live Video Streams on Mobile Devices
Modeling Pixel Process with Scale Invariant Local Patterns for Background Subtraction in Complex Scenes (CVPR’10) Shengcai Liao, Guoying Zhao, Vili Kellokumpu,
Active Calibration of Cameras: Theory and Implementation Anup Basu Sung Huh CPSC 643 Individual Presentation II March 4 th,
Robust Object Segmentation Using Adaptive Thresholding Xiaxi Huang and Nikolaos V. Boulgouris International Conference on Image Processing 2007.
Ensemble Tracking Shai Avidan IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE February 2007.
Human Posture Recognition with Convex Programming Hao Jiang, Ze-Nian Li and Mark S. Drew School of Computing Science Simon Fraser University Burnaby, BC,
Real-time Hand Pose Recognition Using Low- Resolution Depth Images
1 Integration of Background Modeling and Object Tracking Yu-Ting Chen, Chu-Song Chen, Yi-Ping Hung IEEE ICME, 2006.
Smart Traveller with Visual Translator. What is Smart Traveller? Mobile Device which is convenience for a traveller to carry Mobile Device which is convenience.
California Car License Plate Recognition System ZhengHui Hu Advisor: Dr. Kang.
Interactive Sand Art Draw Using RGB-D Sensor Presenter : Senhua Chang.
Face Processing System Presented by: Harvest Jang Group meeting Fall 2002.
Real-Time Face Detection and Tracking Using Multiple Cameras RIT Computer Engineering Senior Design Project John RuppertJustin HnatowJared Holsopple This.
LOGO FACE DETECTION APPLICATION Member: Vu Hoang Dung Vu Ha Linh Le Minh Tung Nguyen Duy Tan Chu Duy Linh Uong Thanh Ngoc CAPSTONE PROJECT Supervisor:
Knowledge Systems Lab JN 8/24/2015 A Method for Temporal Hand Gesture Recognition Joshua R. New Knowledge Systems Laboratory Jacksonville State University.
Distinctive Image Features from Scale-Invariant Keypoints By David G. Lowe, University of British Columbia Presented by: Tim Havinga, Joël van Neerbos.
FEATURE EXTRACTION FOR JAVA CHARACTER RECOGNITION Rudy Adipranata, Liliana, Meiliana Indrawijaya, Gregorius Satia Budhi Informatics Department, Petra Christian.
Professor : Yih-Ran Sheu Student’s name : Nguyen Van Binh Student ID: MA02B203 Kinect camera 1 Southern Taiwan University Department of Electrical Engineering.
Introduction Kinect for Xbox 360, referred to as Kinect, is developed by Microsoft, used in Xbox 360 video game console and Windows PCs peripheral equipment.
Information Extraction from Cricket Videos Syed Ahsan Ishtiaque Kumar Srijan.
3D Fingertip and Palm Tracking in Depth Image Sequences
Knowledge Systems Lab JN 9/10/2002 Computer Vision: Gesture Recognition from Images Joshua R. New Knowledge Systems Laboratory Jacksonville State University.
Zhengyou Zhang Microsoft Research Digital Object Identifier: /MMUL Publication Year: 2012, Page(s): Professor: Yih-Ran Sheu Student.
Real-time object tracking using Kalman filter Siddharth Verma P.hD. Candidate Mechanical Engineering.
Speaker : Meng-Shun Su Adviser : Chih-Hung Lin Ten-Chuan Hsiao Ten-Chuan Hsiao Date : 2010/01/26 ©2010 STUT. CSIE. Multimedia and Information Security.
A Method for Hand Gesture Recognition Jaya Shukla Department of Computer Science Shiv Nadar University Gautam Budh Nagar, India Ashutosh Dwivedi.
Video Based Palmprint Recognition Chhaya Methani and Anoop M. Namboodiri Center for Visual Information Technology International Institute of Information.
出處: Signal Processing and Communications Applications, 2006 IEEE 作者: Asanterabi Malima, Erol Ozgur, and Miijdat Cetin 2015/10/251 指導教授:張財榮 學生:陳建宏 學號: M97G0209.
資訊工程系智慧型系統實驗室 iLab 南台科技大學 1 A Static Hand Gesture Recognition Algorithm Using K- Mean Based Radial Basis Function Neural Network 作者 :Dipak Kumar Ghosh,
A New Fingertip Detection and Tracking Algorithm and Its Application on Writing-in-the-air System The th International Congress on Image and Signal.
Online Kinect Handwritten Digit Recognition Based on Dynamic Time Warping and Support Vector Machine Journal of Information & Computational Science, 2015.
ECE 8443 – Pattern Recognition EE 3512 – Signals: Continuous and Discrete Objectives: Spectrograms Revisited Feature Extraction Filter Bank Analysis EEG.
Motion Analysis using Optical flow CIS750 Presentation Student: Wan Wang Prof: Longin Jan Latecki Spring 2003 CIS Dept of Temple.
1 Artificial Intelligence: Vision Stages of analysis Low level vision Surfaces and distance Object Matching.
Efficient Visual Object Tracking with Online Nearest Neighbor Classifier Many slides adapt from Steve Gu.
Autonomous Robots Vision © Manfred Huber 2014.
CONTENT 1. Introduction to Kinect 2. Some Libraries for Kinect 3. Implement 4. Conclusion & Future works 1.
Looking at people and Image-based Localisation Roberto Cipolla Department of Engineering Research team
Team Members Ming-Chun Chang Lungisa Matshoba Steven Preston Supervisors Dr James Gain Dr Patrick Marais.
Course14 Dynamic Vision. Biological vision can cope with changing world Moving and changing objects Change illumination Change View-point.
Video Surveillance Under The Guidance of Smt. D.Neelima M.Tech., Asst. Professor Submitted by G. Subrahmanyam Roll No: 10021F0013 M.C.A.
A Recognition Method of Restricted Hand Shapes in Still Image and Moving Image Hand Shapes in Still Image and Moving Image as a Man-Machine Interface Speaker.
Robot Vision.
Over the recent years, computer vision has started to play a significant role in the Human Computer Interaction (HCI). With efficient object tracking.
CIRP Annals - Manufacturing Technology 60 (2011) 1–4 Augmented assembly technologies based on 3D bare-hand interaction S.K. Ong (2)*, Z.B. Wang Mechanical.
Creative Coding & the New Kinect
Coding Approaches for End-to-End 3D TV Systems
Kinect for Creative Development with open source frameworks
RGB-D Image for Scene Recognition by Jiaqi Guo
Interactive Visual System
Weihong Li, Hao Tang and Zhigang Zhu
Sign Language Recognition With Unsupervised Feature Learning
Presentation transcript:

Real-Time Hand Gesture Recognition with Kinect for Playing Racing Video Games 2014 International Joint Conference on Neural Networks (IJCNN) July 6-11, 2014, Beijing, China Yanmin Zhu and Bo Yuan Intelligent Computing Lab, Division of Informatics Graduate School at Shenzhen, Tsinghua University Shenzhen , P. R. China

outline Introduction Kinect and Experiment Environment Hand Gesture Recognition Technologies Applications Conclusion

Introduction Hand gesture recognition takes a significant position in the field of human-computer interaction (HCI) hand gesture recognition has been applied to a wide range of interactive systems, such as video games, virtual reality, sign language recognition and robot control.

Introduction Two method : glove-based hand gesture recognition and vision based hand gesture recognition vision based hand gesture recognition is more appealing to users due to its natural and intuitive operation However, vision based hand gesture recognition is confronted with many tough challenges such as cluttered background and illumination change

Hard point 1.complex background with regions similar to skin color can interfere with the segmentation of hand 2.the change in illumination can alter the appearance of hand due to the influence of shadows most traditional detection, tracking and recognition techniques with high precision are computationally intensive and cannot satisfy this demanding constraint(real-time).

Introduction the introduction of RGB-D sensors such as Kinect and ASUS Xtion in recent years has significantly simplified the process of vision based object recognition, especially the segmentation phase These sensors provide remarkable improvement when dealing with complex backgrounds and the depth data can also provide more information about the object and help improve the performance of recognition RGB-D sensors are becoming more and more popular in computer vision and pattern recognition

Introduction This paper can track the movements of both hands and recognize both dynamic gestures and poses of hands For one-hand gesture recognition, users can have full control of the slide show in PowerPoint without using keyboard and mouse For two-hand gesture recognition, users can smoothly control the popular racing video game Need for Speed, which is very demanding in response time

The prominent features 1.our system can deal with strongly cluttered background such as the presence of multiple people and there is no need for users to wear marking devices 2. our approach has comparatively low computational complexity with fast response speed and its practical performance is good enough to be used with high end video games

outline Introduction Kinect and Experiment Environment Hand Gesture Recognition Technologies Applications Conclusion

Kinect Kinect consists of an RGB camera, an IR emitter, an IR depth sensor, a microphone array and a tilt The RGB camera can capture three-channel data in a 1280 × 960 resolution at 12 FPS or a 640 × 480 resolution at 30 FPS The depth map’s resolution can be 320 × 240 or 640 × 480, both at 30 FPS.

Kinect The angular field of view is 57 degrees horizontally and 43 degrees vertically Kinect can be tilted up or down by 27 degrees the valid operating distance of Kinect is 0.8m~4m

Experiment Environment we used both the RGB video stream and the depth map produced by the Kinect sensor with a resolution of 640 × 480 at 30 FPS The computer system was a standard PC with Intel Core i CPU at 3.0 GHz and 8G RAM The OS was 64-bit Windows 7 and our system was developed using Microsoft Visual C++ and OpenCV we used Microsoft Office PowerPoint 2010 and Need for Speed: Most Wanted (2012)

outline Introduction Kinect and Experiment Environment Hand Gesture Recognition Technologies Applications Conclusion

Hand Gesture Recognition Technologies

Preprocessing Reason : Poor lighting condition can have a severe influence on skin color detection in RGB images, we need color balance before detecting skin color There are many color balancing algorithms: Scale by Max, Gray World, White World (Perfect Reflector) etc We compared empirically several color balancing approaches and found that Gray World had relative better performance

Preprocessing Fig. 3. The effect of color balancing: (a)original image (b)skin detection result of original image (c)balanced image (d)skin detection result of balanced image

First Hand Detection

1) Locate target candidates : range of [minDepthi, minDepthi + thresh1] 2) Find the connected region of candidates : range [minDepthi, minDepthi + thresh2] 3) Determine the target : Compare the area of each candidate with the area of its corresponding connected region. If the ratio is larger than a threshold

Second Hand Detection It is very natural for a user to put both hands forward with the same pose before he/she starts to perform two-hand gestures, we assume that this is the initial gesture it is assumed that the second hand is on the same depth plane and has similar pose as the first hand. So the depth and shape information can be used to detect the second hand and start tracking both hands

Tracking tracking-by-detection region of interest (ROI) s is set to be half of the circumstance of hand’s bounding box thresh is a fixed threshold based on the longest distance that hand can move in 1/30 second

Tracking Only the depth images are used for tracking for the sake of efficiency smooth surface such as glass and edge of objects. These bugs may occur frequently and interrupt the tracking process. if the hand is missing and the depth of the closest object is significantly different from the hand in the last frame, this object will be removed and a second detection is conducted to increase the robustness of tracking.

Palm and Fist Classification To distinguish palm and fist, we collected a dataset with 500 images of palm and fist and manually segmented them from the background SVM has better performance when dealing with small datasets, especially in nonlinear and high dimensional situations

Gesture Recognition- directional Since the focus of our system is on efficiency, simple methods are preferred 1. we extract part of the trajectory with a certain length, for example 0.5 second, which measures the last 15 frames of the complete trajectory 2. a few key points are sampled from the trajectory. Our system samples three key points: the starting point, the end point and the middle point

Gesture Recognition- directional 3. these key points are matched with those on the individual template

Gesture Recognition- circular 1. as drawing a circle takes more time than directional gestures, the last 1 second part of the trajectory (30 frames) is extracted 2. try to find a closed curve in the trajectory by searching the 30 frames to see whether there is a point coincides with the end point of the trajectory 3. identify four key points of the closed curve with equal time intervals 4. determine if the closed curve forms a circle by imposing restrictions on the distance between these key points 5. according to the sequential relationship of these key points, determine whether the circle is in clockwise or anticlockwise

Gesture Recognition After the recognition of a gesture, the system waits for one second before recognizing another gesture. This is to leave a time interval for the user to prepare for the next gesture, which can reduce the recognition error caused by hand reset

outline Introduction Kinect and Experiment Environment Hand Gesture Recognition Technologies Applications Conclusion

Applications-PowerPoint

Applications

Conclusion The most prominent features of our system are its efficiency and robustness In the future, we will further enhance the function of our system to handle two-hand trajectory recognition and investigate challenging issues such as hand occlusion