0.25
0.5
0.75
1.25
1.5
1.75
2
Let the Shape Speak - Discriminative Face Alignment using Conjugate Priors
Published on Oct 09, 20122990 Views
This work presents a novel Bayesian formulation for aligning faces in unseen images. Our approach is closely related to Constrained Local Models (CLM) and Active Shape Models (ASM), where an ensembl
Related categories
Chapter list
Let the Shape Speak - Discriminative Face Alignment using Conjugate Priors00:00
Demo Video00:03
Introduction00:38
Related Work - Parametric Image Alignment01:21
The Alignment Goal02:46
The Likelihood Term02:55
Local Landmark Detectors03:41
Local Detectors04:14
Local Landmark Detectors - MOSSE Filters04:53
Local Optimization Strategies05:39
KDE Demo Video08:04
MAP Global Alignment08:32
The Prior Term (1)09:19
The Prior Term (2)10:47
The Algorithm11:46
Hierarchical Search (BASM-KDE-H)12:40
Evaluation Results13:27
Tracking Performance14:58
Tracking Performance - video15:19
Labeled Faces in the Wild (LFW)16:10
Conclusions16:28