Home
Videos uploaded by user “tsishkou”
Ferrari - Computer Vision Eyetrack study (Ferrari vs. Lamborghini Clip)
 
01:50
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=ZeBd_F2Bz5Y . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 738 tsishkou
New York City - Human Attention Heatmap, Directions an Fixation Time. Computer Vision Eyetracking study (This Is New York City Clip)
 
01:01
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=W31mr2wonos Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Three Visual Attention centers and their directions are painted in (white-first/green-second and third). Fixation time is given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 565 tsishkou
Face Recognition and Video Indexing for Media Assets Management
 
02:47
Robust (10 false alarms per hour) and Accurate (70% detection rate per image) Face Detection for Video Indexing (various genres) media assets management. For more information please contact tsishkoudv at gmail.com.
Views: 2141 tsishkou
Ronaldinho - Computer Vision Eyetrack study (Ronaldinho: Touch of Gold Clip)
 
01:50
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=lsO6D1rwrKc . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 2544 tsishkou
Google Street View - Computer Vision Eyetracking study (Google Maps Street View from Google Channel)
 
01:44
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=91wuBqlny50 . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 3167 tsishkou
Palm Detection and Recognition
 
01:01
palm detection and recognition for hand gesture recognition application. Very fast (30 fps), accurate and robust (99%). For more information please contact tsishkoudv at gmail.com.
Views: 6427 tsishkou
iPhone - Computer Vision Eyetracking study (Macworld 2007- Steve Jobs introduces iPhone)
 
02:50
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=s_f-KK140vM . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 4559 tsishkou
Automatic Road Marking Recognition for Highway Data Services
 
01:52
Automatic Road Marking Recognition for highway asset management, mapping and street level imaging. Very fast (100 fps), accurate and robust. For more information please contact tsishkoudv at gmail.com.
Views: 4788 tsishkou
Computer Vision based Eyetracking and Heatmap Generation of Websites - Google, Flickr, Photo Sharing
 
02:28
Automatic Human Visual Attention Modeling for Personalized Web Services. AdSence Heatmap generation using computer vision only. Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 100000 training (human-based) samples. Very fast processing time - less than 1 minute per web page. Accurate and robust (80% compared to eye tracking results on the same database). Tracking of visitors. Automatic website eye gaze tracking and optimization using computer vision only. For more information please contact tsishkoudv at gmail.com.
Views: 12007 tsishkou
Automatic Road Surface Recognition for Highway Data Services
 
02:45
Automatic Road Surface Recognition for highway asset management, mapping and street level imaging. Very fast (100 fps), accurate and robust. For more information please contact tsishkoudv at gmail.com.
Views: 1966 tsishkou
Sony - Computer Vision Eyetrack study (Sony Bravia Ad)
 
02:31
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=oP5J4W5GQ3w . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 1990 tsishkou
Olympics 2008 - Human Attention Heatmap, Directions an Fixation Time. Computer Vision Eyetracking study (Olympics 2008 - Beijing welcome u full version)
 
06:51
Automatic Human Visual Attention Modeling for Video. Human Vision model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=1HEndNYVhZo Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Three Visual Attention centers and their directions are painted in (white-first/green-second and third). Fixation time is given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 3944 tsishkou
Michael Jordan - Human Attention Heatmap, Directions an Fixation Time. Computer Vision Eyetracking study (Top 10 Amazing Michael Jordan Clip)
 
03:50
2FPS. Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=-WmNHGA5jh8 Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Three Visual Attention centers and their directions are painted in (white-first/green-second and third). Fixation time is given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 6109 tsishkou
Human vs Computer! Real Eyetrack study vs Computer vision algorithm  (Heineken Tapvat Clip)
 
00:28
Human Visual Attention model (right) vs. real eyetracking study (left - floating yellow dot/circles). Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Have a look at original videos here http://www.youtube.com/watch?v=TT3O2lUQTUE and here http://www.youtube.com/watch?v=sZWsuEljCMo . Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 3396 tsishkou
Animation - Eyetracking and Heatmaps Generation in Video using Computer Vision - Dragonball AF
 
00:36
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=SUkGZFBuwPs . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 652 tsishkou
Generic Object Recognition in Video
 
01:48
Object recognition for advertisement counting applications. Real-time, accurate and robust. For more information please contact tsishkoudv at gmail.com.
Views: 3218 tsishkou
TV - Real-Time Advertisement Integration into Live Video - Formula 1
 
09:19
Real time integration of advertisement into live broadcasting. Computer vision-based. Ability to use the same media content with different advertisements which are specific to particular audience (geo, hotel, ...). For more information please contact tsishkoudv at gmail.com.
Views: 2142 tsishkou
Automatic Ad Integration into Video
 
01:10
Automatic advertisement integration into video using computer vision based eyetracking technology. Full control over level of human attention that would be spread between the brand and the video! Database of advertisements can be matched to video clips, so each pair would have the best visual co-appearance. Fast (0.5 sec per frame), Accurate (90% accuracy tested on 1200 human-verified image samples) and completely automatic. Other applications include automatic annotations, image+text http://www.youtube.com/watch?v=mPeo6ZR1iYs for more information please contact tsishkoudv at gmail.com
Views: 265 tsishkou
Supermarket - Eyetrack and Heatmap using Computer Vision based Human Visual Attention model vs Real Eyetracking study - catalogue Carrefour
 
01:23
Human Visual Attention model-based computer vision algorithm's generated eyetracking heatmap vs real one from eyetrack study (white cross) http://www.youtube.com/watch?v=o4d8UKNWyvo . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 5545 tsishkou
F1 on PS3 - Human Attention Heatmap, Primary Attention Center. Computer Vision Eyetracking study (Formula One Ps3 - Pure realism !! Clip)
 
04:09
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=NLzA28Z3rhU Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Primary visual Attention center is painted in white. Fixations are given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 930 tsishkou
Automatic Ad Integration into Images
 
01:23
Automatic advertisement integration into images using computer vision based eyetracking technology. Full control over level of human attention that would be spread between the brand and the image! Database of advertisements can be matched to images, so each pair would have the best visual co-appearance. Fast (0.5 sec), Accurate (90% accuracy tested on 1200 human-verified samples) and completely automatic. Other applications include automatic annotations, image+text http://www.youtube.com/watch?v=mPeo6ZR1iYs for more information please contact tsishkoudv at gmail.com
Views: 283 tsishkou
Wii - Computer Vision Eyetrack study (Wii for All Commercial)
 
02:02
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=p5cPVP_llfo . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 710 tsishkou
Crysis Physics - Human Attention Heatmap, Primary Attention Center. Computer Vision Eyetracking study (Crysis Physics - 3'000 barrel explosion made with SandBox 2 Clip)
 
04:00
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=YG5qDeWHNmk Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Primary visual Attention center is painted in white. Fixations are given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 1695 tsishkou
Nike - Computer Vision Eyetrack study (Nike: Take It To the NEXT LEVEL)
 
02:01
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=anwlpTgbQTE . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 1658 tsishkou
Apple - Computer Vision Eyetrack study (1984 Apple's Macintosh Commercial)
 
00:59
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=OYecfV3ubP8 . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 771 tsishkou
Webpage to Video Converter! News
 
09:03
Top webnews from digg.com on your iPhone via YouTube! Completely automatic video news generation service. Computer vision based eyetracking is used to combine image and text so you can see them both at the same time. For more information please contact tsishkoudv at gmail.com. If you'd like any information from your site which have been included in news2video to be removed, please contact me. I process removals at publishers' request.
Views: 961 tsishkou
Heineken - Computer Vision Eyetracking study (Heineken Tapvat Clip)
 
00:30
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=sZWsuEljCMo . Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 954 tsishkou
Martini - Computer Vision vs Real Eyetrack study (Martini Commercial)
 
00:31
Human Visual Attention model vs. real eyetracking study (floating blue dot/circle). Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Have a look at original video here http://www.youtube.com/watch?v=S5qM5iGJLdI . Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 4968 tsishkou
Day - Night Eyetracking simulation using computer vision
 
00:33
day night computer vision-based eyetracking and simulation
Views: 246 tsishkou
NFL vs NHL - Human Attention Heatmap, Directions an Fixation Time. Computer Vision Eyetracking study (NFL vs NHL - Which one is the most brutal game? Clip)
 
02:07
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=gnwYhSz2P5E Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Three Visual Attention centers and their directions are painted in (white-first/green-second and third). Fixation time is given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 1704 tsishkou
Euro 2008 - Eyetracking and Heatmaps Generation in Video using Computer Vision - Euro 2008
 
05:21
2FPS. Automatic Human Visual Attention Modeling for Video. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 2497 tsishkou
Virtual Face Generation
 
00:20
Fast (50000 views per second) Virtual Face Generation (from a single frontal face image) with application to face recognition problem. For more information please contact tsishkoudv at gmail.com.
Views: 3468 tsishkou
Overhang Objects Detection and Measurement for Highway Data Services
 
01:38
Overhang Objects Detection and Measurement for highway asset management, mapping and street level imaging. Very fast (100 fps), accurate and robust. For more information please contact tsishkoudv at gmail.com.
Views: 986 tsishkou
People Counter and Unicity Tailgate Detection - Video Surveillance for Banks
 
01:31
Very accurate (99.99% per trial) and very robust (0.01% per trial) people counter and unicity tailgate detector for video surveillance applications. Very useful to automatically distinguish between one or multiple people in indoor installations. Works even if two persons are very close to each other (robbery, hostages). Specially done for bank security purposes. It uses low resolution top view video camera and low-end PC for real time processing and automatic generation of alarms. For more information please contact tsishkoudv at gmail.com.
Views: 5502 tsishkou
Second Life - Human Attention Heatmap, Primary Attention Center. Computer Vision Eyetracking study (Introduction to Second Life Clip)
 
04:14
Automatic Human Visual Attention Modeling for Video. Human Visual Attention model-based eyetracking study. Have a look at original video here http://www.youtube.com/watch?v=b72CvvMuD6Q Example of an eyetracking "heatmap" that shows how much users are likely to look at different parts of a video. Areas where users looked the most are colored red-black; the yellow-green areas indicate fewer fixations, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Primary visual Attention center is painted in white. Fixations are given for fast video scanning mode. Automatic video eyetracking and advertisement optimization using computer vision only. Heatmap generation using computer vision only. No expensive eye tracking studies needed. Both low level (color, edge, texture, ...) and high level (text recognition and understanding, object recognition and understanding, ) analysis is done. Direction of human attention is computed using 80000 training (human-based) samples. Very fast processing time - 2 fps. Accurate and robust (87% compared to eye tracking results on the same database). For more information please contact tsishkoudv at gmail.com.
Views: 912 tsishkou
Automatic Bridge Recognition and Height Measurement for Highway Data Services
 
02:14
Computer vision-based automatic recognition of bridges and measurement of their height for highway asset management, mapping and street level imaging. Very fast (100 fps), accurate and robust. For more information please contact tsishkoudv at gmail.com.
Views: 2312 tsishkou