Artificial intelligence based approach for object detection in digital image processing

Abstract

During the recent times, the rising capabilities of Computer Vision Technology has revolutionised acquisition of data, its management, modelling, and forecasting using advanced analytics. The innovations taking place in the computer vision domain has been interdisciplinary, in which, the various image processing techniques, and the Artificial Intelligence (AI) based models along with their implementation have played a key role. The current technical advances have empowered the scholars to perform the computational trials, which are otherwise quite difficult to execute with the existing conventional methods. Nowadays, Object Detection (OD) has been a popular field of research amongst scholars. Object detection is a method used for detecting the occurrences of objects within images/videos. The objective of these methods is to replicate human intelligence when it comes to quickly identifying the objects from images or videos. These methods mostly utilize algorithms from Machine learning based approaches to provide the output. The definitive objective of object detection is to locate key items, draw rectangular bounding boxes around them, and define the class of each item identified. Multiple applications exist in area of object detection such as face detection and recognition, Optical Character Recognition (OCR), driver less cars, video surveillance systems, pedestrian detection, and medical imaging, including traffic signs detection and recognition. There are multiple approaches used by researchers for object detection. These approaches use machine learning and deep learning algorithms. Both approaches have diverse ways of execution. Though the object detection as a research area is popular among scholars, there remains, substantial challenges that needs to be addressed. Few of the challenges that have been pertinent when it comes to object detection are combined objective of not only classifying the object in the image correctly but also determine its accurate position. There are times when items in image are disp