| Title | : | Handwritten Image Recognition in Low-resource Indic Scripts |
| Speaker | : | Dr.Partha Pratim Roy (Associate Professor, CSE at IIT Roorkee) |
| Details | : | Sat, 23 Sep, 2023 3:00 PM @ MR1 (SSB 233) |
| Abstract: | : | Text image understanding has long been an active research area because of its complexity and challenges due to a variety of shapes. The deep learning-based text image analytics tasks have shown impressive results under the setting of full supervision of completely labeled datasets. However, the creation of such datasets with a large volume of samples is a challenging and time-consuming task for some unexplored scripts, especially Indic scripts. This research presentation will highlight a few solutions towards effective analysis of textual image analysis in scarcity of data annotation. One of the frameworks exploits a large resource of dataset for training and uses it for cross-language learning of low-resource scripts. To introduce unseen variations of data in deep neural network models, we propose an Adversarial Feature Deformation Module (AFDM) that learns ways to elastically warp extracted features in a scalable manner. The AFDM is inserted between intermediate layers and trained alternatively with the original framework, boosting its capability to better learn highly informative features rather than trivial ones. We record results for varying sizes of training data and observe that our enhanced network generalizes much better in the low-data regime. |
