Yunsheng Zhang

TL;DR: In this paper, a method of segmenting a moving target in a non-restrictive scene based on a full convolution network is proposed, which belongs to the technical field of video object segmentation, and the method comprises steps: firstly, framing is carried out on the video, and a result after framing is used for making a Ground Truth set S for a sample image; a deep feature estimator for an image foreground target is acquired, target maximum intra-class likelihood mapping information in all frames is obtained, and initial prediction on the foreground and the background in

TL;DR: This work represents the wavelet-domain hidden Markov tree model of the original image as a structured energy function, which is proved convex in marginal distributions, and derives the maximum lower bound of the energy function through Lagrange dual transform for the purpose of incorporating marginal constraints into optimization.

TL;DR: A novel approach to improve localization accuracy by fusing multiscale information of the tree model is presented and energy function is generated by combining boundary term estimated by classification likelihoods with regional term obtained by approximation coefficients.

TL;DR: After analyzing the architecture and characteristics of control system in process industry, the multi-agent technology was introduced into process industry process control, and a hierarchical distributed multi- agent control system was proposed.

TL;DR: In this paper, a method for unsupervised segmentation of complex targets from a dynamic scene based on multi-scale combination feature convex optimization is proposed, where wavelet domain, spatial domain and time domain edge features are fused to obtain a multiscale combined probability edge response.