The use of data mining and artificial intelligence technology in art colors and graph and images of computer vision under 6G internet of things communication

To further optimize the existing methods in the field of computer vision and improve the intelligence of image data mining technology, the relevant feedback technology is combined with traditional image data mining technology, and an image data mining technology based on the relevant feedback K-Nearest-Neighbor (K-NN) algorithm is designed and further optimized. The focus is the actual feature extraction test for image color and shape. The test results reveal that for retrieval of K = 1, K = 2, K = 3 images, both positive feedback K-NN algorithm and negative feedback K-NN algorithm can effectively improve the accuracy of image data mining. Among them, negative feedback K-NN algorithm has the highest accuracy for image shape feature extraction. When there are K = 3 images, the accuracy of image data mining can reach 78.3%. Then, the image mining research is conducted on multiple databases. In a total of four databases, the accuracy of image retrieval increases with the increase of feedback times. At the same time, using the optimized KNN algorithm can greatly improve the accuracy of image feature extraction, and the highest accuracy can reach 99.3%. The research content can provide a scientific reference for the follow-up study of KNN algorithm.