Scene-adaptive Moving Detection with Machine Learning Based on Clustering

Moving detection is a hot research area. Order to detect flexible and improve the detection accuracy, we propose a scene-adaptive moving detection model with machine learning based on clustering. The model uses a group of testing images to train the camera firstly, which means we can get out accurate parameters for one scene. We design a detection algorithm that used in training process based on clustering. Then it uses the parameters and detection algorithm to detect the changes in monitor scene. The experience shows our model has a high adaptability and accuracy.