MF-OMKT: Model fusion based on online mutual knowledge transfer for breast cancer histopathological image classification.

Pathological diagnosis is considered as the benchmark for the detection of breast cancer. With the increasing number of patients, computer-aided histopathological image classification can assist pathologists in improving breast cancer diagnosis accuracy and working efficiency. However, a single model is insufficient for effective diagnosis, and this also does not accord with the principle of centralized decision-making. Starting from the real pathological diagnosis scenario, we propose a novel model fusion framework based on online mutual knowledge transfer (MF-OMKT) for breast cancer histopathological image classification. The OMKT part based on deep mutual learning (DML) imitates the mutual communication and learning between multiple experienced pathologists, which can break the isolation of single models and provides sufficient complementarity among heterogeneous networks for MF. The MF part based on adaptive feature fusion uses the complementarity to train a powerful fusion classifier. MF imitates the centralized decision-making process of these pathologists. We used the MF-OMKT model to classify breast cancer histopathological images (BreakHis dataset) into benign and malignant as well as eight subtypes. The accuracy of our model reaches the range of [99.27 %, 99.84 %] for binary classification. And that for multi-class classification reaches the range of [96.14 %, 97.53 %]. Additionally, MF-OMKT is applied to the classification of skin cancer images (ISIC 2018 dataset) and achieves an accuracy of 94.90 %. MF-OMKT is an effective and versatile framework for medical image classification.

TiM-Net: Transformer in M-Net for Retinal Vessel Segmentation.

retinal image is a crucial window for the clinical observation of cardiovascular, cerebrovascular, or other correlated diseases. Retinal vessel segmentation is of great benefit to the clinical diagnosis. Recently, the convolutional neural network (CNN) has become a dominant method in the retinal vessel segmentation field, especially the U-shaped CNN models. However, the conventional encoder in CNN is vulnerable to noisy interference, and the long-rang relationship in fundus images has not been fully utilized. In this paper, we propose a novel model called Transformer in M-Net (TiM-Net) based on M-Net, diverse attention mechanisms, and weighted side output layers to efficaciously perform retinal vessel segmentation. First, to alleviate the effects of noise, a dual-attention mechanism based on channel and spatial is designed. Then the self-attention mechanism in Transformer is introduced into skip connection to re-encode features and model the long-range relationship explicitly. Finally, a weighted SideOut layer is proposed for better utilization of the features from each side layer. Extensive experiments are conducted on three public data sets to show the effectiveness and robustness of our TiM-Net compared with the state-of-the-art baselines. Both quantitative and qualitative results prove its clinical practicality. Moreover, variants of TiM-Net also achieve competitive performance, demonstrating its scalability and generalization ability. The code of our model is available at https://github.com/ZX-ECJTU/TiM-Net.

DCML: Deep contrastive mutual learning for COVID-19 recognition.

COVID-19 is a form of disease triggered by a new strain of coronavirus. Automatic COVID-19 recognition using computer-aided methods is beneficial for speeding up diagnosis efficiency. Current researches usually focus on a deeper or wider neural network for COVID-19 recognition. And the implicit contrastive relationship between different samples has not been fully explored. To address these problems, we propose a novel model, called deep contrastive mutual learning (DCML), to diagnose COVID-19 more effectively. A multi-way data augmentation strategy based on Fast AutoAugment (FAA) was employed to enrich the original training dataset, which helps reduce the risk of overfitting. Then, we incorporated the popular contrastive learning idea into the conventional deep mutual learning (DML) framework to mine the relationship between diverse samples and created more discriminative image features through a new adaptive model fusion method. Experimental results on three public datasets demonstrate that the DCML model outperforms other state-of-the-art baselines. More importantly, DCML is easier to reproduce and relatively efficient, strengthening its high practicality.

[Effect of tongbian navel paste on colonic motility in children with constipation of slow transmission type].

OBJECTIVE: To observe the therapeutic effect of Tongbian Navel Paste (TBP, a self-formulated preparation consisting of both Chinese herbal and Western medicines) on children with constipation of slow transmission type (CSTT) and it influence on patients' colonic motility. METHODS: Sixty-eight children with CSTT were randomly assigned to two groups, 38 in the treatment group treated with TBP, and 30 in the control group treated with oral taking Maren Zipi pill. The changes of clinical symptoms, and the outcomes of colon transmission test were observed and compared before and after treatment. RESULTS: Colon transmission test showed the 48 h and 72 h barium discharging rate (%) in the treatment group before treatment was 10.1 +/- 3.2 and 46.2 +/- 3.9; after treatment, it raised to 59.9 +/- 4.1 and 73.6 +/- 3.6 respectively (P <0.05). The total effective rate in the treatment group was 89.47% and 73.33% in the control group, the difference between groups was significant (P<0.05). CONCLUSION: TBP could promote the motility of colon, it is a safe, convenient and effective preparation for treatment of CSTT.