Image segmentation is a fundamental building block of automatic medical applications. It has been greatly improved since the emergence of deep neural networks. However, deep-learning-based models often require a large amount of manual annotations, which has seriously hindered their practical usage. To alleviate this problem, numerous works were proposed by utilizing unlabeled data based on semi-supervised frameworks. Recently, the Mean-Teacher (MT) model has been successfully applied in many scenarios due to its effective learning strategy. Nevertheless, the existing MT model still has certain limitations. Firstly, to gain extra generalization ability through consistency training, various sorts of perturbations are often added to the training data. However, if the variation is too weak, it may cause the Lazy Student Phenomenon, and cause fluctuations in the learning model. On the contrary, large image perturbations may enlarge the performance gap between the teacher and student. In this case, the student may lose its learning momentum, and more seriously, drag down the overall performance of the whole system. In order to address these issues, we introduce a novel semi-supervised medical image segmentation framework, in which a Cross-Mix Teaching paradigm is proposed to provide extra data flexibility, thus effectively avoid Lazy Student Phenomenon. Moreover, a lightweight Transductive Monitor is applied to serve as the bridge that connects the teacher and student for active knowledge distillation. In the light of this cross-network information mixing and transfer mechanism, our method is able to continuously explore the discriminative information contained in unlabeled data. Extensive experiments on challenging medical image data sets demonstrate that our method outperforms current state-of-the-art semi-supervised segmentation methods under severe lack of supervision.

The evaluation of clinical status and vital signs is a crucial component of remote medical care. Radar provides a non-contact monitoring measurement without consideration of lighting or privacy. However, the multipath effect generated by multiple people gathered in a small indoor space seriously affects the detection of vital sign information. This study presents a non-contact measurement of the vital signs based on stepped-frequency continuous wave ultra-wideband(SFCW-UWB) and impulse radio ultra-wideband(IR-UWB) radars. Additionally, this study suggests the Spectral Unmixing Successive Variational Mode Decomposition (SUSVMD) algorithm since the respiratory components have a probability of mixing with the heartbeat signal’s frequency. The suggested approach extracts band-limited modes through back projection and spectral unmixing optimization. The alternate direction method of multiplication is used to address the optimization problem. The experiments are conducted under two radar measurements and three subjects. The results show that the suggested method is superior and effective.

The accurate segmentation of brain tissue in Magnetic Resonance Image (MRI) slices is essential for assessing neurological conditions and brain diseases. However, it is challenging to segment MRI slices because of the low contrast between different brain tissues and the partial volume effect. 2-Dimensional (2-D) convolutional networks cannot handle such volumetric image data well because they overlook spatial information between MRI slices. Although 3-Dimensional (3-D) convolutions capture volumetric spatial information, they have not been fully exploited to enhance representative ability of deep networks; moreover, they may lead to overfitting with insufficient training data. In this paper, we propose a novel convolutional mechanism, termed Rubik convolution, to capture multi-dimensional information between MRI slices. Rubik convolution rotates the axis of a set of consecutive slices, enabling 2-D convolution kernels to extract features of each axial plane simultaneously. Next, feature maps are rotated back to fuse multidimensional information by the Max-View-Maps. Furthermore, we propose an efficient 2-D convolutional network, namely Rubik-Net, where the residual connections and the bottleneck structure are used to enhance information transmission and reduce the number of network parameters. The Rubik-Net shows promising results on iSeg2017, iSeg2019, IBSR and BrainWeb datasets in terms of segmentation accuracy. In particular, we achieved the best results in 95th percentile Hausdorff distance and average surface distance in cerebrospinal fluid segmentation on the most challenging iSeg2019 dataset. The experiments indicate that Rubik-Net improves the accuracy and efficiency of medical image segmentation. Moreover, Rubik convolution can be easily embedded into existing 2-D convolutional networks.

Hand gesture recognition with radar sensors is essential because they can detect gestures despite environmental factors like lighting, dust, and complex backgrounds. Considering the complexity of a system, it is challenging to design CNNs on CPU devices and realize the carry-on mid-air gesture recognition. We propose a mid-air gesture recognition method based on a novel discriminant feature, and it be used as part of a measurement system of hand movements using an ultrawideband (UWB) radar. The Gesture-ProxylessNAS (GPNAS) is presented to enhance the adaptability of model search and overcome the challenge of the network's computational complexity. In order to fully extract local spatial discriminant features and prevent information loss, Local Binary Pattern (LBP) encoders are utilized to extract local spatial information. In the meantime, multi-layer ShuffleNet with depthwise separable convolution is used to gradually leverage high-level spatial features. The Gesture-ProxylessNAS module revisits the multi-layer ShuffleNet's design spaces using an optimization problem, greatly reducing the network's parameters and computational complexity. According to experimental verification on real UWB hand gestures, the proposed framework provides more satisfactory recognition performance and efficiency with a deeper network structure and fewer parameters. The proposed hand gesture recognition system can recognize gestures with a promising accuracy of 96.52% on the UWB-gestures public dataset [1].

With the development of technology, using radar for gesture recognition is feasible and valuable. However, ensuring that gesture recognition can be applied to a wide range of scenarios with sufficient accuracy is still challenging. Due to the lack of accuracy and efficiency of traditional methods, we propose a gesture recognition scheme based on deep learning. We converted radar signals into pictures and designed a lightweight network called self-reparameterization network based on distance and velocity aware and binary coding(SR-DVBNet) to match them. We use the Self-reparameterization Encoder of the signal as the baseline of the network and add Distance and Velocity-aware Embedding (DVAE) between different blocks to do weighting for different dimensions. Since gesture recognition by radar signals often uses two-dimensional data, such as RDM or chirp-arranged matrices, we designed the DVAE module, which can weigh the different dimensions of the data separately to enhance the interpretability and gesture recognition accuracy of the model. At the same time, we use binary descriptors as the final representation of feature vectors for classification, which can well reflect the features of images and improve classification accuracy. Finally, we verify the algorithm’s effectiveness on two publicly available data sets and achieve an accuracy rate of more than 98%, surpassing other known gesture recognition algorithms.