0.25
0.5
0.75
1.25
1.5
1.75
2
Challenges in Learning the Appearance of Faces for Automated Image Analysis - Part 1
Published on Feb 25, 20073889 Views
The variability of images of the human face challenges research in machine vision since its beginning. Sources of variability not only include individual appearance but also cover external parameters
Chapter list
Challenges in Learning the Appearance of Faces for Automated Image Analysis: part I 00:03
actually, i’m gonna talk about:01:38
the problem(s)03:00
where we are: face detection06:35
face recognition and the like 08:03
motivation08:57
some approaches09:27
SVM: global detector (Poggio’s group)12:49
SVM: component-based detector (Poggio’s group)15:36
global vs component-based19:03
naive bayes (Kanade’s group)20:05
AdaBoost (Viola & Jones)23:39
summing up27:45
what we do31:27
one possible way to object detection 33:57
testing hypotheses34:34
CBCL database 35:29
training by hypothesis testing 36:01
image measurements37:34
ranklets (Smeraldi, 2002)38:01
vertical ranklets (variance-to-natural support ratio)38:06
a salient and a non-salient feature38:21
independent feature selection38:45
independent feature selection38:53
testing38:55
multiple tests39:27
some results (franceschi et al, 2004) 472 positive vs 23,573 negatives39:29
once you have detected a face…41:05