Open set domain adaptation method and system based on entropy minimization

Abstract

The invention provides an open set domain adaptation method based on entropy minimization. The open set domain adaptation method comprises the steps of creating an open set domain adaptation network, and initializing network hyper-parameters; inputting a source domain picture and a target domain picture into an open set domain adaptive network, obtaining picture depth features through a networkfeature extraction layer, and then obtaining prediction probability vectors of various categories through a softmax layer according to a feature map; calculating an entropy loss function by using a known category prediction probability vector and a picture label so as to correctly identify a common category; on the basis, calculating binary cross entropy and a binary diversity loss function by using an unknown category prediction probability value so as to correctly classify unknown categories, updating network parameters by using a back propagation gradient until a loss function value is minimum, and stopping training the network; and storing the network model and the training result, and introducing the target domain data set into the network model to obtain a final target domain label.

Description

Technical field [0001] The invention relates to a neural network learning method, in particular to an open set domain adaptation method and system based on entropy minimization. Background technique [0002] In the past few years, domain adaptation has attracted more and more attention because large-scale labeling of large-scale data sets is usually expensive. Through domain adaptation, the impact of the difference between the target domain and the source domain is reduced, and a fully labeled data set (source domain) is available. [0003] Most domain adaptation methods assume that the source and target domains share the categories of their samples. However, in many practical situations, these methods are usually not feasible, where unknown samples may appear in the target domain. Therefore, the concept of open-domain adaptation is introduced, which is obviously different from the traditional concept of closed-domain adaptation. The goal of open set domain adaptation is to class...

AI Technical Summary

Problems solved by technology

[0003] Most domain adaptation methods assume that the source and target domains share the categories of their samples, however, these methods are often not feasible in many practical situations where unknown samples may appear in the target domain

Images