Performance Enhancement Leveraging Mask-RCNN on Bengali Document Layout Analysis

TL;DR

This paper improves Bengali document layout analysis by adapting Mask R-CNN with hyperparameter tuning, achieving a high dice score and highlighting language-specific challenges in model transfer.

Contribution

It presents a tailored Mask R-CNN approach for Bengali documents, demonstrating the importance of language-specific training and hyperparameter optimization.

Findings

Achieved a dice score of 0.889 on BaDLAD dataset.

Language-specific models outperform transfer learning from English.

Hyperparameter tuning significantly improved model performance.

Abstract

Understanding digital documents is like solving a puzzle, especially historical ones. Document Layout Analysis (DLA) helps with this puzzle by dividing documents into sections like paragraphs, images, and tables. This is crucial for machines to read and understand these documents. In the DL Sprint 2.0 competition, we worked on understanding Bangla documents. We used a dataset called BaDLAD with lots of examples. We trained a special model called Mask R-CNN to help with this understanding. We made this model better by step-by-step hyperparameter tuning, and we achieved a good dice score of 0.889. However, not everything went perfectly. We tried using a model trained for English documents, but it didn't fit well with Bangla. This showed us that each language has its own challenges. Our solution for the DL Sprint 2.0 is publicly available at https://www.kaggle.com/competitions/dlsprint2/discussion/432201 along with notebooks, weights, and inference notebook.