Design and Optimization of Big Data Image Recognition System Based on Deep Learning

This paper aims to optimize image recognition technology in the context of big data and proposes a deep learning algorithm combining convolutional neural network (CNN) and graph convolution network (GCN). Through a well-designed experimental framework, the CNN-GCN fusion algorithm first uses CNN to extract the local features of the image and then uses GCN to capture the global structure information of the image, thus realizing the effective fusion of features. Experimental results show that the accuracy, recall, and F1 score of this algorithm are better than those of traditional CNN, Recurrent Neural Network and GCN models and that it performs well in a fine-grained classification and robustness test, which proves its strong generalization ability and anti-interference ability. The research results verify the effectiveness of the proposed algorithm and provide a new direction for the further development of image recognition technology.