ABSTRACT

YOLO (You Only Look Once) reframe object detection as a single regression problem to spatially separated bounding boxes and associated class probabilities. In this project, we re-implemented the state-of-the-art YOLO approach in Python, and loaded the pre-trained weights, followed by further training using VOC 2007 train/val dataset. We evaluated the model with VOC 2007 test dataset and achieved 52% mean average precision (mAP). We've tried some parameters and figure out the best range of them. The final results and the video/photo demos are shown in this website.

Figure 1. YOLO Structure

Figure 2. Examples of YOLO object detection