Abstract:

Visually impaired individuals face significant challenges in autonomous mobility and situational awareness. Most existing assistive technologies address isolated tasks, such as object recognition or text reading, while failing to capture broader environmental context. This work addresses this limitation by proposing a scene-sensitive, low-cost assistive system that delivers holistic situational information. We present Vocal-Eyes, an intelligent smart glasses platform that provides periodic audio descriptions of the surrounding environment. The system employs a cloud-based neural processing pipeline in which visual features are extracted using a Transformer-based architecture. Relational context is modeled through scene graph generation, and scene graphs are translated into natural language via a graph-to-text module. A lightweight hardware prototype captures visual data locally, while computationally intensive processing is offloaded to the cloud to reduce power consumption. The experimental results show that relational, scene-based narration produces more coherent and informative descriptions than object-centric approaches while maintaining acceptable periodic latency. Cost analysis further indicates that Vocal-Eyes is significantly more affordable than comparable commercial smart glasses solutions. These results demonstrate that Transformer-based scene understanding with cloud-assisted processing is an effective and practical approach for developing accessible, context-aware assistive technologies for visually impaired users.
Source code and results at GitHub .

Weakly Supervised Optic Disc and Cup Segmentation Using Pseudo Labels for Automated Glaucoma Analysis.

UnderReview in (AI) MDPI JOUNRAL

Abstract:

Glaucoma is a leading cause of irreversible blindness worldwide, where early detection relies heavily on morphological assessment of the optic nerve head, particularly the vertical Cup-to-Disc Ratio (vCDR). Although deep learning-based segmentation methods have demonstrated strong performance, their dependence on expert pixel-level annotations limits scalability and clinical deployment. In this work, we propose an anatomically guided weakly supervised framework for optic disc and optic cup segmentation that eliminates the need for manual masks. Anatomically plausible pseudo-labels are automatically generated from unannotated retinal fundus images and filtered using physiological quality control criteria to construct a reliable silver-standard training set.
Source code and results at  GitHub. 

© Copyright 2026 Meesam Abbas.