Paper Info
Title
Multi-view facial landmark detector learned by the Structured Output SVM.
Abstract
We propose a real-time multi-view landmark detector based on Deformable Part Models (DPM). The detector is composed of a mixture of tree based DPMs, each component describing landmark configurations in a specific range of viewing angles. The usage of view specific DPMs allows to capture a large range of poses and to deal with the problem of self-occlusions. Parameters of the detector are learned from annotated examples by the Structured Output Support Vector Machines algorithm. The learning objective is directly related to the performance measure used for detector evaluation. The tree based DPM allows to find a globally optimal landmark configuration by the dynamic programming. We propose a coarse-to-fine search strategy which allows real-time processing by the dynamic programming also on high resolution images. Empirical evaluation on \"in the wild\" images shows that the proposed detector is competitive with the state-of-the-art methods in terms of speed and accuracy yet it keeps the guarantee of finding a globally optimal estimate in contrast to other methods. Real-time multi-view landmark detector independent on initializationDetector composed of a mixture of tree based Deformable Part ModelsLandmark detector learned by the Structured Output Support Vector MachinesCoarse-to-fine strategy to speed up inference based on dynamic programming
Year
DOI
Venue
2016
10.1016/j.imavis.2016.02.004
Image Vision Comput.
Keywords
Field
DocType
Deformable Part Models,Structured output SVM,Facial landmarks detection
Computer vision,Dynamic programming,Pattern recognition,Inference,Computer science,Support vector machine,Artificial intelligence,Landmark,Detector,Speedup
Journal
Volume
Issue
ISSN
47
C
0262-8856
Citations 
PageRank 
References 
3
0.43
35
Authors
5
Name
Order
Citations
PageRank
Michal Uricar1864.31
Vojtěch Franc258455.78
Diego Thomas36711.42
Akihiro Sugimoto438342.87
Václav Hlavác561685.46