Breast cancer is the most common malignancy among women worldwide, and accurate lesion segmentation is of great importance for its early diagnosis and treatment. However, due to the high morphological variability of lesions and the inherent complexity of ultrasound imaging, existing deep-learning-based methods still face significant challenges in achieving satisfactory segmentation accuracy for breast ultrasound images. To address this limitation, this study proposed a novel breast-lesion segmentation network, termed CWSASKM-BBAM-Net, which is built upon the classical U-Net architecture. First, a Channel-Wise Spatially Adaptive Selective Kernel Convolution Module (CWSASKM) was introduced to adaptively adjust the receptive field size for each spatial location based on channel-specific semantic features, thereby enhancing the network’s multi-scale representation learning. Second, a Bidirectional Boundary-Aware Mechanism (BBAM) was designed to jointly model salient regions and their boundaries by integrating forward and reverse attention. This mechanism progressively improves the discrimination between non-salient areas and lesion areas, thereby refining boundary delineation. Extensive experiments were conducted on three public breast ultrasound datasets (BUSI, UDIAT, and STU). On BUSI dataset, the proposed method achieved Jaccard index, precision, recall, and Dice similarity coefficient of 71.97%, 82.85%, 81.40%, and 80.44%, respectively, outperforming the second-best approach by margins of 1.69, 1.05, 1.28, and 1.84 percentage points. On UDIAT, it attained 78.14%, 88.31%, 86.73%, and 86.10% on these metrics, with improvements of 2.75, 2.04, 0.56, and 2.01 percentage points, respectively. The proposed method also demonstrated superior overall performance on the external STU dataset. These results collectively demonstrate that CWSASKM-BBAM-Net achieves state-of-the-art performance in breast ultrasound image segmentation tasks.

Named entity recognition is a fundamental task of natural language processing (NLP) and plays an important role in many downstream NLP tasks, including information extraction and machine translation, etc. The existing named entity recognition methods are usually based on sequence labeling and extract entities within a sentence independently. These methods ignore the semantic information between sentences. Named entity recognition methods based on machine reading comprehension encode important prior information about the entity class. It is easier to distinguish similar classification labels, which reduces the difficulty of model learning, but it still only models at the sentence level, ignoring the semantic information between sentences, which is easy to cause the problem of inconsistent entity labeling in different sentences. To this end, this paper extended the sentence-level named entity recognition to the text-level named entity recognition, and then proposed a BiLSTM-BiDAF named entity recognition model based on machine reading comprehension. First, to utilize the context information within the whole text, NEZHA pre-training language model was used to obtain information of the full text and local features were further captured through BiLSTM, so as to strengthen the model’s ability to capture locally dependent information. Then, a bidirectional attention flow was introduce to learn the semantic association between the text and entity category. Finally, to predict the position of entities in the text, a boundary detector based on the gating mechanism was design to strengthen the correlation of the entity boundary. At the same time, an answer count detector was establish to identify the unanswerable questions. Experimental results on the CCKS2020 Chinese electronic medical records dataset and CMeEE dataset show that our model can effectively identify document-level and sentence-level named entities, and F₁ can reach 84.76% and 57.35%, respectively.