Input/Output Deep Architecture for Structured Output Problems
Preprint
Abstract
Pre-training of input layers has shown to be efficient for learning deep architectures, solving the vanishing gradient issues. In this paper, we propose to extend the use of pre-training to output layers in order to address structured output problems, which are characterized by internal dependencies between the outputs (e.g. the classes of pixels in an image labeling problem). Whereas the output structure is generally modeled using graphical models, we propose a fully neural-based model called IODA (Input Output Deep Architecture) that learns both input and output dependencies. We apply IODA on facial landmark detection problems where the output is a strongly structured regression vector. We evaluate IODA on two public challenging datasets: LFPW and HELEN. We show that IODA outperforms the traditional pre-training approach.Keywords
All Related Versions
This publication has 23 references indexed in Scilit:
- IODA: An input/output deep architecture for image labelingPattern Recognition, 2015
- Face detection, pose estimation, and landmark localization in the wild2012
- Real-time facial feature detection using conditional regression forests2012
- Deformable Model Fitting by Regularized Landmark Mean-ShiftInternational Journal of Computer Vision, 2010
- Enzyme Function Prediction with Interpretable Models2009
- Learning to Localize Objects with Structured Output Regression2008
- Pictorial Structures for Object RecognitionInternational Journal of Computer Vision, 2005
- A statistical approach for phrase location and recognition within a text line: an application to street name recognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2002
- Active Shape Models-Their Training and ApplicationComputer Vision and Image Understanding, 1995
- Auto-association by multilayer perceptrons and singular value decompositionBiological Cybernetics, 1988