Federal learning-based image classification method and device, computer equipment and medium

• Main Authors: LEE TAEG-WON, WANG JIANZONG
• Format: Patent
• Language: Chinese; English
• Published: 14.07.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention relates to the technical field of artificial intelligence, in particular to an image classification method and device based on federated learning, computer equipment and a medium. According to the method, a label-free image is segmented and randomly shielded to obtain a normal sub-image set and a shielded sub-image, an initial model sent by a server is received, the normal sub-image set is input into an encoder to obtain a feature vector, the feature vector and the shielded sub-image are input into a decoder to obtain a reconstructed image, and the reconstructed image is decoded. Training an initial model according to the label-free image and the reconstructed image, sending the trained encoder to a server, forming a classification model by a reference encoder sent by the server and a classifier, training the classification model, inputting the to-be-processed image into the trained classification model, and obtaining an image classification result. According to the method, image reconstruction