Advanced Two Stage AI Technique for Object Detection

Advanced Two Stage AI Technique for Object Detection

Shashank Tiwari*1, Balla Sahithi*2, M. Phani Krishna*3, M. Hari Charan Reddy*4

*1 Assistant Professor Of Department Of CSE (AI & ML) Of ACE Engineering College, India.

*2,3,4 Students Of Department Of CSE (AI & ML) Of ACE Engineering College, India.

ABSTRACT

Object detection in computer vision uses AI, mainly deep learning, to identify and locate objects in images or videos. It involves an AI system that spots various objects, determines their type, and marks their positions with bounding boxes. Built on advanced deep learning models like Convolutional Neural Networks (CNNs), YOLO, or Faster R-CNN, it excels in real-time detection. Trained on large datasets like COCO, it recognizes diverse objects across different scenes. It tackles challenges investigates challenges like varying object sizes, scales, lighting, and occlusion using techniques such as transfer learning, data augmentation, and model optimization for fast, accurate results.

Keywords:

Deep learning, YOLO / Faster R-CNN, bounding boxes, and COCO dataset