Accurate and unbiased examinations of skin lesions are critical for the early diagnosis and treatment of skin diseases. Visual features of skin lesions vary significantly because the images are collected from patients with different lesion colours and morphologies by using dissimilar imaging equipment. Recent studies have reported that ensembled convolutional neural networks (CNNs) are practical to classify the images for early diagnosis of skin disorders. However, the practical use of these ensembled CNNs is limited as these networks are heavyweight and inadequate for processing contextual information. Although lightweight networks (e.g., MobileNetV3 and EfficientNet) were developed to achieve parameter reduction for implementing deep neural networks on mobile devices, insufficient depth of feature representation restricts the performance. To address the existing limitations, we develop a new lite and effective neural network, namely HierAttn. The HierAttn applies a novel deep supervision strategy to learn the local and global features by using multi-stage and multi-branch attention mechanisms with only one training loss. The efficacy of HierAttn was evaluated by using the dermoscopy images dataset ISIC2019 and smartphone photos dataset PAD-UFES-20 (PAD2020). The experimental results show that HierAttn achieves the best accuracy and area under the curve (AUC) among the state-of-the-art lightweight networks. The code is available at https://github.com/anthonyweidai/HierAttn .

Early detection and accurate diagnosis can predict the risk of malignant disease transformation, thereby increasing the probability of effective treatment. Identifying mild syndrome with small pathological regions serves as an ominous warning and is fundamental in the early diagnosis of diseases. While deep learning algorithms, particularly convolutional neural networks (CNNs), have shown promise in segmenting medical objects, analyzing small areas in medical images remains challenging. This difficulty arises due to information losses and compression defects from convolutional and pooling operations in CNNs, which become more pronounced as the network deepens, especially for small medical objects. To address these challenges, we propose a novel scale-variant attention-based network (SvANet) for accurately segmenting small-scale objects in medical images. The SvANet consists of scale-variant attention (SvAttn), cross-scale guidance, Monte Carlo attention (MCAttn), and Vision Transformer (ViT), which incorporates cross-scale features and alleviates compression artifacts for enhancing the discrimination of small medical objects. Quantitative experimental results demonstrate the superior performance of SvANet, achieving 96.12%, 96.11%, 89.79%, 84.15%, 80.25%, 73.05%, and 72.58% in mean Dice (mDice) coefficient for segmenting kidney tumors, skin lesions, hepatic tumors, polyps, surgical excision cells, retinal vasculatures, and sperms, which occupy less than 1% of the image areas in KiTS23, ISIC 2018, ATLAS, PolypGen, TissueNet, FIVES, and SpermHealth datasets, respectively.

Flexible sensors have the advantage to enable intelligent systems to interact with the environment with high safety and versatility. Liquid metal is a useful element for flexible sensors due to its stretchable, reconfigurable, and healing characteristics. However, it is difficult to decode the stimuli position using uniform liquid metal because of its homogeneity and fluidity. Current sensing systems for the perception of force magnitude and contact location mostly rely on sensing arrays and massive crossover electrodes. In this paper, we proposed a novel method for simultaneously sensing the force magnitude and contact location without crossover electrodes. In the new design, we encapsulate the liquid metal electrodes with varying cross-sections in two silicon layers that function as protective skin and govern the geometry of liquid metals. When a force is applied to the flexible sensor, the change of resistance is dependent on both location and stimuli magnitude. Three liquid metal channels with different gradient resistance are embedded in the prototypes with parallel distributions. Combining the signals collected from the three channels, we can simultaneously measure the force magnitude, stimuli location, and contacting orientation. The overall force resolution is sub-Newton in the range from 0.5 N to 20 N. The position can also be determined with a 100% success rate for the segmented sensor and a position resolution of 2 mm for the continuous sensor. The error for contact orientation measurement is 4.58° when the orientation ranges from 45° to 135°. The proposed multifunctional sensing device with high accuracy and wide sensing range has significant potential for developing interactive interfaces for human-machine interactions and industrial applications.