- An Obstacle Avoidance and Person Identification project using Raspberry Pi, semantic segmentation, and deep learning to detect obstacles and identify individuals. Trained on ADE20K and COCO datasets, it provides real-time audio feedback to visually impaired users. - Facial recognition was done in the engineering firm. The dataset was collected from the faces of the employees. There were four images for each person, and then the image augmentation was applied, making it around 16 images… Show more - An Obstacle Avoidance and Person Identification project using Raspberry Pi, semantic segmentation, and deep learning to detect obstacles and identify individuals. Trained on ADE20K and COCO datasets, it provides real-time audio feedback to visually impaired users. - Facial recognition was done in the engineering firm. The dataset was collected from the faces of the employees. There were four images for each person, and then the image augmentation was applied, making it around 16 images per person, which helped get better accuracy. Show less