Analytics of Deep Neural Network-Based Background Subtraction

Abstract

Deep neural network-based (DNN-based) background subtraction has demonstrated excellent performance for moving object detection. The DNN-based background subtraction automatically learns the background features from training images and outperforms conventional background modeling based on handcraft features. However, previous works fail to detail why DNNs work well for change detection. This discussion helps to understand the potential of DNNs in background subtraction and to improve DNNs. In this paper, we observe feature maps in all layers of a DNN used in our investigation directly. The DNN provides feature maps with the same resolution as that of the input image. These feature maps help to analyze DNN behaviors because feature maps and the input image can be simultaneously compared. Furthermore, we analyzed important filters for the detection accuracy by removing specific filters from the trained DNN. From the experiments, we found that the DNN consists of subtraction operations in convolutional layers and thresholding operations in bias layers and scene-specific filters are generated to suppress false positives from dynamic backgrounds. In addition, we discuss the characteristics and issues of the DNN based on our observation.

Robust Vehicle Detection and Counting Algorithm Employing a Convolution Neural Network and Optical Flow

“…Recently [3,4,16,17], convolutional neural networks (ConvNets) have presented excellent results in different vision challenges where it has shown an attractive characteristic to learn deep and hierarchical features, which make it more powerful than classical methods. In this work, two convolution layers, two max-pooling layers, and two fully connected feed-forward layers are adopted with the same network architecture in [16], which obtained better detection results by discriminating the foreground and background regions.…”

Deep detector classifier (DeepDC) for moving objects segmentation and classification in video surveillance

“…The network takes a red–blue–green (RGB) image in three different scales and generates a foreground segmentation probability mask for the corresponding image. In the period of 2018–2019, numerous deep learning models either based on auto‐encoder [29–31] and CNNs [32–36] have been proposed. However, all these methods are supervised and have been trained on ground truth video frames of datasets and tested on the same types of videos.…”

Semi-Supervised Faster RCNN-Based Person Detection and Load Classification for Far Field Video Surveillance

“…The next step of the approach consists of passing the detected moving areas to a person detector to generate boxes around the person in the scene. In our previous work [22], the person detection was that can be used to find moving areas-for example, see the literature [24][25][26][27][28][29][30][31][32][33]-a simple frame differencing is utilized here to provide the input to the deep learning-based person detector. The person detector then corrects remaining errors associated with moving areas.…”