Research on Image Recognition System Based on Computer Artificial Intelligence Technology

This paper uses mobile network mode to develop and model Python crawler based on Ubuntu16.04 and CUDA9.0. The visualization of this pattern was achieved using Tensor Board software. The hardware design of the system consists of three modules: linear array CCD sensor, DSP processor and signal processing circuit. Through the reconstruction and learning of such models, the trained models are stored, so as to obtain higher precision image recognition. A method of classifying objects using Fourier description operators is proposed. The experimental results show that the accuracy of the image is 96.8% and the recognition time is 0.55 seconds. Motion tracking is based on the positioning and capture of the specific position of the human hand in each screen, and the captured frame rate is 28 frames per second. The system can quickly acquire images of 1024*1500 pixels at a rate of 5 photos/second. Through the static identification of posture and tracking of motion trajectory, the interaction between human and machine is better.