Video Object Detection Using Densenet-Ssd
2021
Despite the fact that there are well-established object detection methods that are based on images, the application of these approaches to video data on a frame-by-frame basis has two major drawbacks.One is inefficiency in terms of processing caused either by redundant information across image frames or by the absence of temporal and spatial correlation of features across image frames.And other is not being able to cope with the challenges of everyday utilization, such as motion blur and occlusion.In order to capture both temporal and spatial correlation of features, we proposed an improved SSD algorithm called DenseNet -SSD.From the experimental analysis, it is explained that our proposed object detection model gives better result.