Edge detection based boundary box construction algorithm for improving the precision of object detection in YOLOv3

In the field of computer vision, object detection is one of the most interesting field, and one major task after object has been detected is to draw boundary boxes around detected object. The proposed work focuses on improving the precision of boundary boxes that had been drawn after objects has being detected on image. When it comes to object detection YOLOv3 is a real time state-of-the-art object detection system. Due to this, the proposed work had taken YOLOv3 as base model and improves the precision of the boundary boxes around the object. The proposed framework uses pretrained COCO dataset for object detection, and with the help of edge detection and pixel values in an area, the proposed work improves the precision of the boundary box around the object. The proposed model gives significantly better precision of boundary boxes when compared with YOLOv3.