Joint Cascade Face Detection and Alignment
JDA a new state-of-the-art approach for face detection. The key idea is to combine face alignment with detection, observing that aligned face shapes provide better features for face classification. To make this combination more effective, the approach learns the two tasks jointly in the same cascade framework, by exploiting recent advances in face alignment. Such joint learning greatly enhances the capability of cascade detection and still retains its realtime performance. Extensive experiments show that our approach achieves the best accuracy on challenging datasets, where all existing solutions are either inaccurate or too slow.
The frame work is as follows:
The train process:
The result:
The detect result is 78% on FDDB, which is not as good as the paper, this is mainly due the training samples are not as rich as the paper uses.
But on the speed side, I get the 60 FPS on a normal VGA image. which is much faster than the paper itself.
But on the speed side, I get the 60 FPS on a normal VGA image. which is much faster than the paper itself.