AUTHOR=Chen Yehang , Chen Xiangmeng 

TITLE=A brain-like classification method for computed tomography images based on adaptive feature matching dual-source domain heterogeneous transfer learning

JOURNAL=Frontiers in Human Neuroscience

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2022.1019564

DOI=10.3389/fnhum.2022.1019564

ISSN=1662-5161

ABSTRACT=Transfer learning can improve the robustness of deep learning in the case of small samples. However, when the semantic difference between the source domain data and the target domain data is large, transfer learning easily introduces redundant features and leads to negative transfer. Therefore, a classification method based on adaptive feature matching dual-source domain heterogeneous transfer learning was proposed for the preoperative aided diagnosis of lung granuloma and lung adenocarcinoma for patients with solitary pulmonary solid nodule in the case of small samples. The method includes two parts: 1) feature extraction and 2) feature classification. In the feature extraction part, first, an adaptive selected-based dual-source domain feature matching network was proposed to determine the matching weight of each pair of feature maps between the two source networks and the target network and the matching weight of each pair of convolutional layers between the two source networks and the target network. These two weights can adaptively select the features in the source network that are conducive to the learning of the target task and the destination of feature transfer to constrain the training of the target network and improve the robustness of the target network in the case of small samples. Meanwhile, a target network based on diverse branch block was proposed, which made the target network have different receptive fields and complex paths to further improve the feature expression ability of the target network. Second, the convolution kernel of the target network was used as the feature extractor to extract features. In the feature classification part, an ensemble classifier based on sparse Bayesian extreme learning machine was proposed that can automatically decide how to combine the output of base classifiers to improve the classification performance. Finally, an experiment was carried out on the data of two medical centers. The experimental results (the AUCs were 0.9542 and 0.9356, respectively) showed that this method can provide a better diagnostic reference for doctors.