Browsing by Author "Kumar, K.S.C."

Now showing 1 - 6 of 6

A strategy for selecting optimal vergence pose and calibration patterns for stereo-vision calibration
(Institute of Electrical and Electronics Engineers Inc., 2016) Kumar, K.S.C.
The usage of stereo-vision in Automobile, Automation, machine-vision, Infotainment etc. is growing these days enormously day by day. To design a stereo-vision camera, we need to know optimal vergence pose between two cameras as well as number of stereo-pairs required for calibration so that stereo-vision can yield dense disparity map, precise 3D reconstruction, higher stereo-calibration accuracy as well as lower calibration time. This paper addresses these problems of stereo-vision prior to calibration by selecting optimal vergence pose between two cameras and optimal number of stereo-calibration pattern pairs for calibration. To estimate optimal vergence pose, a set of virtual stereo-calibrations patterns are generated by varying Euler angles and it is observed that stereo-reprojection error decreases as Roll angle reduces and exhibits mirror image property and is less sensitive to the variations in Yaw and Pitch. Also to estimate the optimal number of stereo-calibration pattern pairs, an epipolar constraint error was used in eliminating redundant pairs thereby reducing stereo-calibration time and lower stereo-reprojection error. Â© 2015 IEEE.
Phase-edge based approach for pedestrian segmentation using NIR camera and tracking for driver assistance
(2013) Kumar, K.S.C.
A new approach using phase edge based techniquefor night time pedestrian segmentation and tracking using Near IR camera is proposed in this paper. A Gabor filter is used to extract even weak potential vertical edges that belong to candidate pedestrian blocks even under non-uniform illuminations and poor contrast. Vertical and horizontal projection on Gabor filtered response is carried out to determine the bounding-box of pedestrian-like candidate blobs. Pedestrian like candidate blocks undergo through a series of rule based classifiers. Once such set of pedestrian blocks are identified, candidate pedestrian blocks are tracked in next consecutive frames of a video using density confidence based tracker which uses Gabor filter edge response for feature-space analysis. False positives are greatly removed using spatio-temporal analysis and tracking mechanism. Performance characterization of the algorithm has been carried out in highway roads under nonuniform illuminations and varying contrast. Experimental result shows that algorithm can produce a detection rate of nearly 69% and can be used in real-time. � 2013 IEEE.
Phase-edge based approach for pedestrian segmentation using NIR camera and tracking for driver assistance
(2013) Kumar, K.S.C.
A new approach using phase edge based techniquefor night time pedestrian segmentation and tracking using Near IR camera is proposed in this paper. A Gabor filter is used to extract even weak potential vertical edges that belong to candidate pedestrian blocks even under non-uniform illuminations and poor contrast. Vertical and horizontal projection on Gabor filtered response is carried out to determine the bounding-box of pedestrian-like candidate blobs. Pedestrian like candidate blocks undergo through a series of rule based classifiers. Once such set of pedestrian blocks are identified, candidate pedestrian blocks are tracked in next consecutive frames of a video using density confidence based tracker which uses Gabor filter edge response for feature-space analysis. False positives are greatly removed using spatio-temporal analysis and tracking mechanism. Performance characterization of the algorithm has been carried out in highway roads under nonuniform illuminations and varying contrast. Experimental result shows that algorithm can produce a detection rate of nearly 69% and can be used in real-time. Â© 2013 IEEE.
Stereo-vision based smart TV control
(2016) Kumar, K.S.C.
With the rapid advances in the field of computer vision systems and increasing flexibility in using multimedia applications, brings major challenges in terms of algorithm design and hardware. One such widely used infotainment application is Smart TV control in Multimedia portfolio. This paper presents a non-intrusive approach for smart TV control virtual assistance system targeting elderly citizens using stereo-vision by switching TV ON/OFF and volume INR/DECR events without using any hand-held remote control. System is initialized by waving the hands in-view of stereo vision mounted on Television (TV). Hand palm regions are detected and extracted using motion and color-space model. In next successive frames of a video, Hand palms are tracked using Color Histogram based tracker. A virtual 3D plane is then constructed by waving hand palms around in view of stereo-vision camera and reconstructing 3D points using Epipolar geometry and Triangulation. A KEY PRESS/RELEASE(TV ON/OFF) events are identified by measuring the distance between projected 3D forefinger tip into/onto virtual 3D plane and TV VOLUME INCR/DECR events are identified by waving the hands UP/DOWN in-view of stereo-vision. � 2015 IEEE.
Stereo-vision based smart TV control
(Institute of Electrical and Electronics Engineers Inc., 2016) Kumar, K.S.C.
With the rapid advances in the field of computer vision systems and increasing flexibility in using multimedia applications, brings major challenges in terms of algorithm design and hardware. One such widely used infotainment application is Smart TV control in Multimedia portfolio. This paper presents a non-intrusive approach for smart TV control virtual assistance system targeting elderly citizens using stereo-vision by switching TV ON/OFF and volume INR/DECR events without using any hand-held remote control. System is initialized by waving the hands in-view of stereo vision mounted on Television (TV). Hand palm regions are detected and extracted using motion and color-space model. In next successive frames of a video, Hand palms are tracked using Color Histogram based tracker. A virtual 3D plane is then constructed by waving hand palms around in view of stereo-vision camera and reconstructing 3D points using Epipolar geometry and Triangulation. A KEY PRESS/RELEASE(TV ON/OFF) events are identified by measuring the distance between projected 3D forefinger tip into/onto virtual 3D plane and TV VOLUME INCR/DECR events are identified by waving the hands UP/DOWN in-view of stereo-vision. Â© 2015 IEEE.
A strategy for selecting optimal vergence pose and calibration patterns for stereo-vision calibration
(2016) Kumar, K.S.C.
The usage of stereo-vision in Automobile, Automation, machine-vision, Infotainment etc. is growing these days enormously day by day. To design a stereo-vision camera, we need to know optimal vergence pose between two cameras as well as number of stereo-pairs required for calibration so that stereo-vision can yield dense disparity map, precise 3D reconstruction, higher stereo-calibration accuracy as well as lower calibration time. This paper addresses these problems of stereo-vision prior to calibration by selecting optimal vergence pose between two cameras and optimal number of stereo-calibration pattern pairs for calibration. To estimate optimal vergence pose, a set of virtual stereo-calibrations patterns are generated by varying Euler angles and it is observed that stereo-reprojection error decreases as Roll angle reduces and exhibits mirror image property and is less sensitive to the variations in Yaw and Pitch. Also to estimate the optimal number of stereo-calibration pattern pairs, an epipolar constraint error was used in eliminating redundant pairs thereby reducing stereo-calibration time and lower stereo-reprojection error. � 2015 IEEE.