An efficient deep Convolutional neural network approach for object detection and recognition from a video sequence using multi scale anchor box

Abstract

Deep learning is a new era of machine learning which trains computers to find the newlinestructure from a massive amount of data. Learning is described at multiple levels of newlinerepresentation. This enables us to make sense of the data consisting of text, sound, newlineand images. Many computer vision problems such as object detection, image newlineclassification, and semantic segmentation have been solved using convolutional newlineneural networks. Object detection in videos involves confirming the presence of the newlineobject in the image or video and then locating it accurately for recognition. Detecting newlineand recognizing the still object from an image has comparatively shown better newlineperformance with the use of detection frameworks like R-CNN, Fast R-CNN, Faster newlineR-CNN etc. The challenge is to detect and recognize the moving object from a static newlinecamera efficiently and accurately. In the video, modeling techniques suffer from high newlinecomputation and memory costs which may lead to a decrease in performance newlinemeasures such as accuracy and efficiency to identify the object accurately. The newlinemotive behind this work is to accurately detect and recognize the moving and still newlineobject from a video sequence using deep learning in real-time. The existing newlinealgorithms of object detection based on the deep convolution neural network worked newlinewell for large-size objects as the detection models get better results. However, those newlinemodels fail to detect the varying size of the objects that have low resolution. This is newlinebecause the features do not fully represent the essential characteristics of the objects newlinein real-time after going through the repeated convolution operations of existing newlinemodels. The proposed work improves the accuracy of detection by extracting the newlinefeatures of object at different size and scale by using a multi-scale anchor box. With newlinethe help of CNN, the deep knowledge from the dataset is extracted by giving the newlinemodel a rigorous set of training samples.