ColorInsightX: A Personalized AI-Powered Mobile Assistive System for Color Vision Deficiency Using Object Detection and Daltonization

Color vision deficiency(CVD) is a disability that affects

millions of people in the world and that has an effect on the perceiver of

color when it comes to distinguishing colors in their daily activities like

drawing clothes and differentiating objects. Although the recent

developments in computer vision made it possible to apply the assistive

technology to color-blind users, many of the existing systems rely on

simple color perception and do not customize and provide context

sensitive input. As introduced in this paper, ColorInsightX is an AI-based

assistive system that offers convenience to people with CVD in terms of

accessibility. The proposed system will combine the Farnsworth Munsell

100 Hue Test to identify the approach of the type of color vision

deficiency and confusion axis of the user, ZOLO-Tiny object detector

which is also based on the YOLO architecture, and color extraction

algorithms to identify the dominant colors. As a visual accessibility

measure, the system also has Daltonization based color correction that

produces custom visuals depending on the users color perception profile.

This system is developed with the help of react native in the front-end and

FastAI in the back-end so that color recognition and object-detection via

camera can be done in real time. There are also text-to-speech feedback

which gives audio feedback on objects and colors found and a rule-based

outfit suggestion module which will aid in the user choosing outfits that

are notable. ColorInsightX provides a context-dependent, individualized,

and multi- modal assistive technology as compared to the conventional

color recognition applications; this enablespeople with color vision

deficiency to gain greater independence and usability