Identification of blurred objects in real time Video using deep learning neural networks
Ayush Senapati, Atul Kumar Karn, Yash Bhardwaj, S. Jenicka, R. Priya, Ankit Shukla, Shivam Arora · 2022 13th International Conference on Computing Communication and Networking Technologies (ICCCNT) · 2022
Our method employs a novel strategy for blur detection. The algorithm has been split into two stages. In the first stage, we used the YOLO object detection algorithm to detect the biggest object in the image. We also trained a Convolutional neural network (CNN) model with an external dataset for blur detection and finetuned it to suit our application. We captured the real-time video and the trained CNN model was able to detect whether the frames were blurred or not.