An Investigation of Deep Neural Network based Techniques for Object Detection and Recognition Task in Computer Vision

An Investigation of Deep Neural Network based Techniques for Object Detection and Recognition Task in Computer Vision

연구 분야: Artificial Intelligence

학회: 2023 2nd International Conference on Edge Computing and Applications (ICECAA)

Detection of objects and its recognition in visual sequences are the two critical tasks in the computer vision field. Various real-time applications such as autonomous vehicles, face recognition, health-care systems and space exploration requires highly reliable and precise object detection models. Traditional object detection and recognition algorithms are based on hand crafted and are considered to be erroneous, time consuming and expensive leading to the significant reduction of accuracy rate for object detection in large datasets. Recently, large number of promising deep neural networks models have been emerged for facilitating automated and accurate detection of varying scale objects and its precise recognition across various computer vision applications. Several GPU based neural models thereby incorporating context-aware capabilities have shown effective performance, which overcomes the drawbacks of traditional techniques. This research study provides an investigation of several popular deep learning models that exists for accurate object detection in various forms of visual sequences. Varying scale objects are detected frequently available in popular MS COCO and PASCAL datasets and their performance are evaluated utilizing one stage Yolo family and two stage Faster RCNN deep learning object detectors. At the end of the study, several future research directions for object detection task are discussed.