An interpretable transformer network for the retinal disease classification using optical coherence tomography.

Retinal illnesses such as age-related macular degeneration and diabetic macular edema will lead to irreversible blindness. With optical coherence tomography (OCT), doctors are able to see cross-sections of the retinal layers and provide patients with a diagnosis. Manual reading of OCT images is time-consuming, labor-intensive and even error-prone. Computer-aided diagnosis algorithms improve efficiency by automatically analyzing and diagnosing retinal OCT images. However, the accuracy and interpretability of these algorithms can be further improved through effective feature extraction, loss optimization and visualization analysis. In this paper, we propose an interpretable Swin-Poly Transformer network for performing automatically retinal OCT image classification. By shifting the window partition, the Swin-Poly Transformer constructs connections between neighboring non-overlapping windows in the previous layer and thus has the flexibility to model multi-scale features. Besides, the Swin-Poly Transformer modifies the importance of polynomial bases to refine cross entropy for better retinal OCT image classification. In addition, the proposed method also provides confidence score maps, assisting medical practitioners to understand the models' decision-making process. Experiments in OCT2017 and OCT-C8 reveal that the proposed method outperforms both the convolutional neural network approach and ViT, with an accuracy of 99.80% and an AUC of 99.99%.

Information bottleneck-based interpretable multitask network for breast cancer classification and segmentation.

Breast cancer is one of the most common causes of death among women worldwide. Early signs of breast cancer can be an abnormality depicted on breast images (e.g., mammography or breast ultrasonography). However, reliable interpretation of breast images requires intensive labor and physicians with extensive experience. Deep learning is evolving breast imaging diagnosis by introducing a second opinion to physicians. However, most deep learning-based breast cancer analysis algorithms lack interpretability because of their black box nature, which means that domain experts cannot understand why the algorithms predict a label. In addition, most deep learning algorithms are formulated as a single-task-based model that ignores correlations between different tasks (e.g., tumor classification and segmentation). In this paper, we propose an interpretable multitask information bottleneck network (MIB-Net) to accomplish simultaneous breast tumor classification and segmentation. MIB-Net maximizes the mutual information between the latent representations and class labels while minimizing information shared by the latent representations and inputs. In contrast from existing models, our MIB-Net generates a contribution score map that offers an interpretable aid for physicians to understand the model's decision-making process. In addition, MIB-Net implements multitask learning and further proposes a dual prior knowledge guidance strategy to enhance deep task correlation. Our evaluations are carried out on three breast image datasets in different modalities. Our results show that the proposed framework is not only able to help physicians better understand the model's decisions but also improve breast tumor classification and segmentation accuracy over representative state-of-the-art models. Our code is available at https://github.com/jxw0810/MIB-Net.

YOLOF-Snake: An Efficient Segmentation Model for Green Object Fruit.

Accurate detection and segmentation of the object fruit is the key part of orchard production measurement and automated picking. Affected by light, weather, and operating angle, it brings new challenges to the efficient and accurate detection and segmentation of the green object fruit under complex orchard backgrounds. For the green fruit segmentation, an efficient YOLOF-snake segmentation model is proposed. First, the ResNet101 structure is adopted as the backbone network to achieve feature extraction of the green object fruit. Then, the C5 feature maps are expanded with receptive fields and the decoder is used for classification and regression. Besides, the center point in the regression box is employed to get a diamond-shaped structure and fed into an additional Deep-snake network, which is adjusted to the contours of the target fruit to achieve fast and accurate segmentation of green fruit. The experimental results show that YOLOF-snake is sensitive to the green fruit, and the segmentation accuracy and efficiency are significantly improved. The proposed model can effectively extend the application of agricultural equipment and provide theoretical references for other fruits and vegetable segmentation.

An Attention Based Bidirectional LSTM Method to Predict the Binding of TCR and Epitope.

The T-cell epitope prediction has always been a long-term challenge in immunoinformatics and bioinformatics. Studying the specific recognition between T-cell receptor (TCR) and peptide-major histocompatibility complex (p-MHC) complexes can help us better understand the immune mechanism, it's also make a signification contribution in developing vaccines and targeted drugs. Meanwhile, more advanced methods are needed for distinguishing TCRs binding from different epitopes. In this paper, we introduce a hybrid model composed of bidirectional long short-term memory networks (BiLSTM), attention and convolutional neural networks (CNN) that can identified the binding of TCRs to epitopes. The BiLSTM can more completely extract amino acid forward and backward information in the sequence, and attention mechanism can focus on amino acids at certain positions from complex sequences to capture the most important feature, then CNN was used to further extract salient features to predict the binding of TCR-epitope. In McPAS dataset, the AUC value (the area under ROC curve) of naive TCR-epitope binding is 0.974 and specific TCR-epitope binding is 0.887. The model has achieved better prediction results than other existing models (TCRGP, ERGO, NetTCR), and some experiments are used to analyze the advantages of our model. The algorithm is available at https://github.com/bijingshu/BiAttCNN.git.

Oral administration of Urtica macrorrhiza Hand.-Mazz. polysaccharides to protect against cyclophosphamide-induced intestinal immunosuppression.

As a strategy to prevent the well-known immunosuppressant effects of cyclophosphamide (CY), the immunomodulatory activity of the polysaccharide isolated from Urtica macrorrhiza Hand.-Mazz. (UMHMPS) was investigated in the present study. The chemical properties of UMHMPS, including total carbohydrates, uronic acid, protein contents, monosaccharide compositions, molecular weight and structural confirmation, were investigated. The immunomodulatory activity of UMHMPS was evaluated using a CY-induced immunosuppression mouse model. The results revealed that UMHMPS, which is composed of rhamnose, gluconic acid, galactose acid, galactose and xylose, exhibited potent immunomodulatory activity and low toxicity in mice. It increased the secretions of secretory immunoglobulin A, interferon (IFN)-γ and interleukin (IL)-4, and maintained the balance of the ratios of IFN-γ/IL-4 and cluster of differentiation (CD)3+/CD19+ cells in Peyer's patches. Furthermore, it increased the expression of Toll-like receptor (TLR)-4, indicating that TLR4 may be one of the receptors of UMHMPS. Therefore, the present study provides evidence for the potential use of UMHMPS as an immune enhancement drug in chemotherapy.