RESUMEN
Breast cancer significantly impacts women's health, with ultrasound being crucial for lesion assessment. To enhance diagnostic accuracy, computer-aided detection (CAD) systems have attracted considerable interest. This study introduces a prospective deep learning architecture called "Multi-modal Multi-task Network" (3MT-Net). 3MT-Net utilizes a combination of clinical data, B-mode, and color Doppler ultrasound. We have designed the AM-CapsNet network, specifically tailored to extract crucial tumor features from ultrasound. To combine clinical data in 3MT-Net, we have employed a cascaded cross-attention to fuse information from three distinct sources. To ensure the preservation of pertinent information during the fusion of high-dimensional and low-dimensional data, we adopt the idea of ensemble learning and design an optimization algorithm to assign weights to different modalities. Eventually, 3MT-Net performs binary classification of benign and malignant lesions as well as pathological subtype classification. In addition, we retrospectively collected data from nine medical centers. To ensure the broad applicability of the 3MT-Net, we created two separate testsets and conducted extensive experiments. Furthermore, a comparative analysis was conducted between 3MT-Net and the industrial-grade CAD product S-detect. The AUC of 3MT-Net surpasses S-Detect by 1.4% to 3.8%.
RESUMEN
Waste pollution is a significant environmental problem worldwide. With the continuous improvement in the living standards of the population and increasing richness of the consumption structure, the amount of domestic waste generated has increased dramatically, and there is an urgent need for further treatment. The rapid development of artificial intelligence has provided an effective solution for automated waste classification. However, the high computational power and complexity of algorithms make convolutional neural networks unsuitable for real-time embedded applications. In this paper, we propose a lightweight network architecture called Focus-RCNet, designed with reference to the sandglass structure of MobileNetV2, which uses deeply separable convolution to extract features from images. The Focus module is introduced to the field of recyclable waste image classification to reduce the dimensionality of features while retaining relevant information. To make the model focus more on waste image features while keeping the number of parameters small, we introduce the SimAM attention mechanism. In addition, knowledge distillation was used to further compress the number of parameters in the model. By training and testing on the TrashNet dataset, the Focus-RCNet model not only achieved an accuracy of 92[Formula: see text] but also showed high deployment mobility.
RESUMEN
Nucleic acid testing is currently the golden reference for coronaviruses (SARS-CoV-2) detection, while the SARS-CoV-2 antigen-detection rapid diagnostic tests (RDT) is an important adjunct. RDT can be widely used in the community or regional screening management as self-test tools and may need to be verified by healthcare authorities. However, manual verification of RDT results is a time-consuming task, and existing object detection algorithms usually suffer from high model complexity and computational effort, making them difficult to deploy. We propose LightR-YOLOv5, a compact rotating SARS-CoV-2 antigen-detection RDT results detector. Firstly, we employ an extremely light-weight L-ShuffleNetV2 network as a feature extraction network with a slight reduction in recognition accuracy. Secondly, we combine semantic and texture features in different layers by judiciously combining and employing GSConv, depth-wise convolution, and other modules, and further employ the NAM attention to locate the RDT result detection region. Furthermore, we propose a new data augmentation approach, Single-Copy-Paste, for increasing data samples for the specific task of RDT result detection while achieving a small improvement in model accuracy. Compared with some mainstream rotating object detection networks, the model size of our LightR-YOLOv5 is only 2.03MB, and it is 12.6%, 6.4%, and 7.3% higher in mAP@.5:.95 metrics compared to RetianNet, FCOS, and R3Det, respectively.