The First Swahili Language Scene Text Detection and Recognition Dataset

TL;DR

This paper addresses the absence of Swahili language resources for scene text detection and recognition by introducing the Swahili-text dataset, comprising 976 natural-scene images with word-level annotations and 8284 cropped word images for recognition. It evaluates six models across detection (DBNet, PANet, FCENet) and recognition (ASTER, SATRN, ABINet), reporting SATRN as the top recognizer (93.9% word-level accuracy) and FCENet as a strong detector (84.8% F-score, with PANet showing competitive metrics). The dataset provides a publicly available benchmark to spur Swahili-language research and highlights the need for language-specific adaptations to achieve optimal performance in real-world scenes. Overall, Swahili-text lays groundwork for future expansion and methodological development in low-resource language scene text understanding.

Abstract

Scene text recognition is essential in many applications, including automated translation, information retrieval, driving assistance, and enhancing accessibility for individuals with visual impairments. Much research has been done to improve the accuracy and performance of scene text detection and recognition models. However, most of this research has been conducted in the most common languages, English and Chinese. There is a significant gap in low-resource languages, especially the Swahili Language. Swahili is widely spoken in East African countries but is still an under-explored language in scene text recognition. No studies have been focused explicitly on Swahili natural scene text detection and recognition, and no dataset for Swahili language scene text detection and recognition is publicly available. We propose a comprehensive dataset of Swahili scene text images and evaluate the dataset on different scene text detection and recognition models. The dataset contains 976 images collected in different places and under various circumstances. Each image has its annotation at the word level. The proposed dataset can also serve as a benchmark dataset specific to the Swahili language for evaluating and comparing different approaches and fostering future research endeavors. The dataset is available on GitHub via this link: https://github.com/FadilaW/Swahili-STR-Dataset

Figures (6)

• Figure 1: Dataset Sample
• Figure 2: Statistical analysis of the Swahili text dataset: (a) Grouping images by text length. (b) Counting character occurrences in the dataset.
• Figure 3: Annotated Image Sample
• Figure 4: Text detection visualization
• Figure 5: Text Recognition Visualizations