Research Of Neural Network Algorithm In Object Pose Estimation And Image Segmentation

With the development of computer science,neural network algorithm has been applied in various fields of social development.The problem of object pose estimation and image threshold segmentation has always been the basic problem in the field of artificial intelligence.In the problem of object pose estimation,it is important to obtain the object category and accurate pose information for the robot to interact with the environment.In the problem of image threshold segmentation,precisely selecting the threshold of image segmentation can better separate the foreground and background of the image,so as to facilitate the next operation of the image.The main research work of this paper is as follows:1.Pose estimation and object detection are important tasks for robot-environment interaction.Impressive progress has been made in this field over the past decade.However,the problem of pose recognition of objects in complex scenes is still challenging.Based on the VGG network,this paper proposes an improved convolutional neural network model that uses a multi-stage approach to extract features from the data.And for a learning-based approach,a lot of data is necessary.Computer using a rapid synthesis method of image data can be generated in a short time a large number of qualified training data.Experimental results demonstrate the effectiveness and better performance of the proposed method by comparing with classical deep neural networks.2.We present a collective neurodynamic optimization algorithm to solve the application of binary optimization in the image segmentation.First,the threshold-based image segmentation problem is transformed into a constrained binary optimization problem.Then a collective neurodynamic optimization algorithm is introduced which combined with feedback neural network and particle swarm optimization(PSO)algorithm.The linear programming relaxation constraint method is used to relax binary constraints.The local convergence of the algorithm is proved by feedback neural network algorithm,better results can be refinement by PSO.Finally,several sets of comparative experiments are presented.The feasibility of our proposed method is verified,the experimental results demonstrate the effectiveness of our approach in image segmentation.