Publications

FAIERDet: Fuzzy-based adaptive image enhancement for real-time traffic sign detection and recognition under varying light conditions

Published in Expert Systems with Applications, 2026

This paper addresses the challenge of recognizing traffic signs in poor visibility under varying light conditions. It employs a two-step approach: find the lighting quality of image and then based on that quality enhance image if quality lies below some threshold. Then employ YOLOv8 to detect the traffic signs.

Download here

Generative Method-Based Traffic Sign Detection and Recognition in Occluded Conditions

Published in IEEE 101st Vehicular Technology Conference (VTC2025-Spring), Oslo, 2025

Priyanka Choudhary and Somnath Dey

This paper proposes a novel feature-based Generative Adversarial Network (GAN) to regenerate occluded traffic signs to improve detection accuracy. The GAN-generated features are fused with the Faster R-CNN-generated features. The architecture of GAN is tiny and integrated at the feature level within Faster R-CNN to minimize computational overhead.

Download here

Automatic Signboard Recognition in Low Quality Night Images

Published in International Conference on Computer Vision and Image Processing (CVIP - 2023), 2023

Manas Kagde, Priyanka Choudhary, Rishi Joshi and Somnath Dey

This paper addresses the challenge of recognizing traffic signs in poor visibility, blurriness, and adverse conditions. It employs a two-step approach: enhancing images with a modified MIRNet model and then recognizing signs using the YOLOv4 model to improve recognition accuracy under difficult conditions.

Download here