Empirical analysis of cascade deformable models for multi-view face detection

作者：

Highlights：

• We present a state-of-the-art multi-view face detector based on Cascade Deformable Part Models (CDPM).

• We propose to combine data-mining and bootstrapping to learn CDPM models from weakly labelled data.

• We report extensive validation of our models in the FDDB, AFLW, HDDB and COFW databases.

• We show the suitability of our models for face alignment initialization and face detection under partial occlusions.

摘要

•We present a state-of-the-art multi-view face detector based on Cascade Deformable Part Models (CDPM).•We propose to combine data-mining and bootstrapping to learn CDPM models from weakly labelled data.•We report extensive validation of our models in the FDDB, AFLW, HDDB and COFW databases.•We show the suitability of our models for face alignment initialization and face detection under partial occlusions.

论文关键词：Multi-view face detection,Cascade deformable models,FDDB database,AFLW database,HPID database,COFW dataset

论文评审过程：Received 11 June 2014, Revised 18 July 2015, Accepted 24 July 2015, Available online 14 August 2015, Version of Record 11 September 2015.

论文官网地址：https://doi.org/10.1016/j.imavis.2015.07.002