Problem
Research questions and friction points this paper is trying to address.
Prosthetic Vision
Image Quality Improvement
Object Recognition
Semantic and structural image segmentation for prosthetic vision
🤖 AI Summary
Low-resolution, achromatic, and low-contrast phosphene images in artificial retinal prostheses hinder scene understanding. To address this challenge—particularly in indoor environments—this paper proposes a semantic-structural joint sketch representation method. We introduce a novel multi-branch collaborative CNN architecture that jointly optimizes structural edge detection and semantic segmentation for the first time, while incorporating phosphene encoding constraints to produce a compact representation preserving both geometric structure and semantic content. Evaluated under simulated prosthetic vision conditions, our method significantly improves object recognition and room classification accuracy over conventional image-processing baselines. It establishes an interpretable and robust perceptual compression paradigm tailored for bandwidth-constrained neural visual prostheses.
📝 Abstract
Prosthetic vision is being applied to partially recover the retinal stimulation of visually impaired people. However, the phosphenic images produced by the implants have very limited information bandwidth due to the poor resolution and lack of color or contrast. The ability of object recognition and scene understanding in real environments is severely restricted for prosthetic users. Computer vision can play a key role to overcome the limitations and to optimize the visual information in the prosthetic vision, improving the amount of information that is presented. We present a new approach to build a schematic representation of indoor environments for simulated phosphene images. The proposed method combines a variety of convolutional neural networks for extracting and conveying relevant information about the scene such as structural informative edges of the environment and silhouettes of segmented objects. Experiments were conducted with normal sighted subjects with a Simulated Prosthetic Vision system. The results show good accuracy for object recognition and room identification tasks for indoor scenes using the proposed approach, compared to other image processing methods.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Image Processing
Semantic Segmentation
Prosthetic User Experience
🔎 Similar Papers
No similar papers found.
