Swin-Net: A Swin-Transformer-Based Network Combing with Multi-Scale Features for Segmentation of Breast Tumor Ultrasound Images.

Breast cancer is one of the most common cancers in the world, especially among women. Breast tumor segmentation is a key step in the identification and localization of the breast tumor region, which has important clinical significance. Inspired by the swin-transformer model with powerful global modeling ability, we propose a semantic segmentation framework named Swin-Net for breast ultrasound images, which combines Transformer and Convolutional Neural Networks (CNNs) to effectively improve the accuracy of breast ultrasound segmentation. Firstly, our model utilizes a swin-transformer encoder with stronger learning ability, which can extract features of images more precisely. In addition, two new modules are introduced in our method, including the feature refinement and enhancement module (RLM) and the hierarchical multi-scale feature fusion module (HFM), given that the influence of ultrasonic image acquisition methods and the characteristics of tumor lesions is difficult to capture. Among them, the RLM module is used to further refine and enhance the feature map learned by the transformer encoder. The HFM module is used to process multi-scale high-level semantic features and low-level details, so as to achieve effective cross-layer feature fusion, suppress noise, and improve model segmentation performance. Experimental results show that Swin-Net performs significantly better than the most advanced methods on the two public benchmark datasets. In particular, it achieves an absolute improvement of 1.4-1.8% on Dice. Additionally, we provide a new dataset of breast ultrasound images on which we test the effect of our model, further demonstrating the validity of our method. In summary, the proposed Swin-Net framework makes significant advancements in breast ultrasound image segmentation, providing valuable exploration for research and applications in this domain.

End-to-end antigenic variant generation for H1N1 influenza HA protein using sequence to sequence models.

The growing risk of new variants of the influenza A virus is the most significant to public health. The risk imposed from new variants may have been lethal, as witnessed in the year 2009. Even though the improvement in predicting antigenicity of influenza viruses has rapidly progressed, few studies employed deep learning methodologies. The most recent literature mostly relied on classification techniques, while a model that generates the HA protein of the antigenic variant is not developed. However, the antigenic pair of influenza virus A can be determined in a laboratory setup, the process needs a tremendous amount of time and labor. Antigenic shift and drift which are caused by changes in surface protein favored the influenza A virus in evading immunity. The high frequency of the minor changes in the surface protein poses a challenge to identifying the antigenic variant of an emerging virus. These changes slow down vaccine selection and the manufacturing process. In this vein, the proposed model could help save the time and efforts exerted to identify the antigenic pair of the influenza virus. The proposed model utilized an end-to-end learning methodology relying on deep sequence-to-sequence architecture to generate the antigenic variant of a given influenza A virus using surface protein. Employing the BLEU score to evaluate the generated HA protein of the antigenic variant of influenza virus A against the actual variant, the proposed model achieved a mean accuracy of 97.57%.

Retinal Vessel Segmentation Based on B-COSFIRE Filters in Fundus Images.

Retinal vessel extraction plays an important role in the diagnosis of several medical pathologies, such as diabetic retinopathy and glaucoma. In this article, we propose an efficient method based on a B-COSFIRE filter to tackle two challenging problems in fundus vessel segmentation: (i) difficulties in improving segmentation performance and time efficiency together and (ii) difficulties in distinguishing the thin vessel from the vessel-like noise. In the proposed method, first, we used contrast limited adaptive histogram equalization (CLAHE) for contrast enhancement, then excerpted region of interest (ROI) by thresholding the luminosity plane of the CIELab version of the original RGB image. We employed a set of B-COSFIRE filters to detect vessels and morphological filters to remove noise. Binary thresholding was used for vessel segmentation. Finally, a post-processing method based on connected domains was used to eliminate unconnected non-vessel pixels and to obtain the final vessel image. Based on the binary vessel map obtained, we attempt to evaluate the performance of the proposed algorithm on three publicly available databases (DRIVE, STARE, and CHASEDB1) of manually labeled images. The proposed method requires little processing time (around 12 s for each image) and results in the average accuracy, sensitivity, and specificity of 0.9604, 0.7339, and 0.9847 for the DRIVE database, and 0.9558, 0.8003, and 0.9705 for the STARE database, respectively. The results demonstrate that the proposed method has potential for use in computer-aided diagnosis.