Computer Vision Aids for Visual Assistance and Guidance

Resumen

Witnessing the growth in the number of visually impaired individuals, the lack of the
availability of electronic travel aid (ETA) devices that are easily accessible becomes a potential
problem for people from low-income backgrounds. This project presents an approach to solve
the common issues visually impaired people face, such as navigation through different
environments and auxiliary information about the objects in their surroundings. A
comprehensive approach towards object detection, distance estimation, and text information
extraction from the user’s environment combined with haptic and auditory feedback attempts
to provide a user-friendly and cost-effective solution using computer vision and deep learning
methods. Feasible application scenarios have also been provided which employ our proposed
methods.