Back to Search
Start Over
CRF with deep class embedding for large scale classification.
- Source :
- Computer Vision & Image Understanding; Feb2020, Vol. 191, pN.PAG-N.PAG, 1p
- Publication Year :
- 2020
-
Abstract
- This paper presents a novel deep learning architecture for classifying structured objects in ultrafine-grained datasets, where classes may not be clearly distinguishable by their appearance but rather by their context. We model sequences of images as linear-chain CRFs, and jointly learn the parameters from both local-visual features and neighboring class information. The visual features are learned by convolutional layers, whereas class-structure information is reparametrized by factorizing the CRF pairwise potential matrix. This forms a context-based semantic similarity space, learned alongside the visual similarities, and dramatically increases the learning capacity of contextual information. This new parametrization, however, forms a highly nonlinear objective function which is challenging to optimize. To overcome this, we develop a novel surrogate likelihood which allows for a local likelihood approximation of the original CRF with integrated batch-normalization. This model overcomes the difficulties of existing CRF methods to learn the contextual relationships thoroughly when there is a large number of classes and the data is sparse. The performance of the proposed method is illustrated on a huge dataset that contains images of retail-store product displays, and shows significantly improved results compared to linear CRF parametrization, unnormalized likelihood optimization, and RNN modeling. We also show improved results on a standard OCR dataset. • A method for training a factorized CRF that enables combining batch-normalization. • Combining deep class embedding into a CRF formulation. • Handling datasets with a huge number of similar classes. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 10773142
- Volume :
- 191
- Database :
- Supplemental Index
- Journal :
- Computer Vision & Image Understanding
- Publication Type :
- Academic Journal
- Accession number :
- 141774749
- Full Text :
- https://doi.org/10.1016/j.cviu.2019.102865