Object detection, localization and tracking using Single Shot Detector (SSD)

Abstract

We are trying to Detect, Localize and classify objects in an image using neural networks. We are using Single Shot detector for performing the task. The output of Single Shot Detector is Bounding boxes. It will try to localize the objects in an image and also tell which object is present in an image. The network predict probability scores of all the classes, we will take the highest probability of all the predicted probabilities. The network also predicts bounding boxes in an image, which can be reduced using NonMax Suppression. Object Detection is most widely used task in computer vision. We can locate objects based on our interest. Object detection is most widely used in real time situations and it has many practical applications. Some of the applications include Vehicle number plate detection, face mask detection, face recognition. Applications include face detection, pedestrian detection and vehicle detection. Single Shot Detector is Deep Learning based algorithm.