OBJECT DETECTION AND LOCALIZATION OF CUSTOM OBJECTS IN A VIDEO

Abstract

The main objective ofthis project is to localize and detect the any moving object in a real time video and obviously that can be one ofthe challenging task in the fields like the computer vision due to several factors such as the centroid or the oiigin of the object, the location ofthe object different types ofshadings in a video to detect etc. By studying the Many ofthe Artificial Intelligence algorithm one ofthe most efficient that we have found is the Convolutional Neural Network or the CNN that has the good accuracy rate and yet the most perfect one to fit for recognizing any custom trained object in a video, A Module will be developed which will further train and detect the custom objects and all the final implementation will be done using the Programming language called ‘Python’ IDE (PYcharm). Just to train a single object there the minimum requirement for a single object is more than the 900 images and if increase the system will get more efficient in detecting that image and by considering that I have noticed that Convolutional Neural Network is the better choice than all other deep learning algorithms there would be one of the training phase and the testing phase as well. Further development is also included in the report. Object Detection is the process of finding real-world object instances like many real life time objects still images or Videos. It allows for the recongnition , localization, and detection ofmultiple objects within an image which provide us with a much understanding of an image as a whole. It is commonly used in applications such as image retrieval, security, surveillance, and advanced driver assistance system (ADAS). ; : ; ! i I : : ; ■■ \ We work on different data set and train them to detect an objects , it give us 100% accuracy result through CNN algorithm.