Multifeature fusion object tracking by radial basis function neural networks

Visual object tracking is an important branch of computer vision. Its main task is to identify and track targets in the scene by analyzing and calculating captured images or videos. In order to track the target effectively and reliably, this paper proposes a multi-feature fusion target tracking algorithm based on RBF neural network. RBF neural network is a simple and efficient three-layer neural network, which can greatly improve tracking efficiency. This paper first uses the three-dimensional color histogram, the Canny operator and the gray-scale image with spatial correlation to construct the input feature vector of the RBF neural network. Then use the triangular kernel function as the activation function of the RBF neural network. Finally, the proposed algorithm is used to track the target. The experimental results show that the target tracking algorithm proposed in this paper can track the target reliably, and has good adaptability to problems such as camera movement, illumination changes, target rotation, and shape deformation.