Improving Image Classification Using Coarse and Fine Labels

Abstract : We consider the problem of image classification using deep convolutional networks, with respect to hierarchical relationships among classes. We investigate if the semantic hierarchy is captured by CNN models or not. For this we analyze the confidence of the model for a category and its sub-categories. Based on the results, we propose an algorithm for improving the model performance at test time by adapting the classifier to each test sample and without any re-training. Secondly, we propose a strategy for merging models for jointly learning two levels of hierarchy. This reduces the total training time as compared to training models separately, and also gives improved classification performance.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01590672
Contributor : Georges Quénot <>
Submitted on : Tuesday, September 19, 2017 - 10:29:08 PM
Last modification on : Tuesday, July 9, 2019 - 1:26:59 AM

Identifiers

Citation

Anuvabh Dutt, Denis Pellerin, Georges Quénot. Improving Image Classification Using Coarse and Fine Labels. Proceedings of the 2017 ACM International Conference on Multimedia Retrieval, Jun 2017, Bucarest, Romania. pp.438--442, ⟨10.1145/3078971.3079042⟩. ⟨hal-01590672⟩

Share

Metrics

Record views

322