LRTCFPan: Low-Rank Tensor Completion Based Framework for Pansharpening.

Pansharpening refers to the fusion of a low spatial-resolution multispectral image with a high spatial-resolution panchromatic image. In this paper, we propose a novel low-rank tensor completion (LRTC)-based framework with some regularizers for multispectral image pansharpening, called LRTCFPan. The tensor completion technique is commonly used for image recovery, but it cannot directly perform the pansharpening or, more generally, the super-resolution problem because of the formulation gap. Different from previous variational methods, we first formulate a pioneering image super-resolution (ISR) degradation model, which equivalently removes the downsampling operator and transforms the tensor completion framework. Under such a framework, the original pansharpening problem is realized by the LRTC-based technique with some deblurring regularizers. From the perspective of regularizer, we further explore a local-similarity-based dynamic detail mapping (DDM) term to more accurately capture the spatial content of the panchromatic image. Moreover, the low-tubal-rank property of multispectral images is investigated, and the low-tubal-rank prior is introduced for better completion and global characterization. To solve the proposed LRTCFPan model, we develop an alternating direction method of multipliers (ADMM)-based algorithm. Comprehensive experiments at reduced-resolution (i.e., simulated) and full-resolution (i.e., real) data exhibit that the LRTCFPan method significantly outperforms other state-of-the-art pansharpening methods. The code is publicly available at: https://github.com/zhongchengwu/code_LRTCFPan.

A long-term survival rat model of spinal cord ischemia injury: Thoracic aortic occlusion combined with aortic bypass circulation.

OBJECTIVE: This study aims to investigate the methods for rat spinal cord ischemia injury models with a high long-term survival rate. METHODS: The rats were divided into three groups: the treatment group, the control group, and the sham operation group. The treatment group had a blocked thoracic aorta (landing zone 3 by Ishimaru - T11) + aortic bypass circulation for 20 min. In the control group, the thoracic aorta at the landing zone 3 was blocked for 20 min. In the sham operation group, only thoracotomy without thoracic aortic occlusion was performed. The mean arterial blood pressure (MABP) of the thoracic aorta and caudal artery before and after thoracic aortic occlusion was monitored intraoperatively. Spinal cord function was monitored by a transcranial motor evoked potential (Tc-MEP) during the operation. Spinal cord function was evaluated by the BBB scale (Basso, Beattie, & Bresnahan locomotor rating scale) scores at multiple postoperative time points. The spinal cord sections of the rats were observed for 7 days after surgery, and the survival curves were analyzed for 28 days after surgery. RESULTS: After aortic occlusion, the MABP of thoracic aorta decreased to 6% of that before occlusion, and the MABP of caudal artery decreased to 63% of that before occlusion in the treatment group. In the control group, the MABP of both thoracic aorta and caudal artery decreased to 19% of that before occlusion. The Tc-MEP waveform of the treatment group disappeared after 6 min, and that of the control group disappeared after 8 min until the end of surgery. There was no change in the Tc-MEP waveform in the sham operation group. The BBB score of the treatment group decreased more obviously than the control group, and there was a significant difference. There was no decrease in the sham group. Spinal cord sections showed a large number of degeneration and necrosis of neurons, infiltration of inflammatory cells, and proliferation of surrounding glial cells in the treatment group. In the control group, multiple neurons were necrotic. The histology of the sham operation group was normal. The 28-day survival rate of the treatment group was 73.3%, which was higher than the control group (40.0%), and there was a significant difference (p < 0.05). CONCLUSION: Thoracic aortic occlusion combined with aortic bypass is an effective modeling method for rats with accurate modeling effects and high long-term survival rates.

Model Lightweighting for Real-time Distraction Detection on Resource-Limited Devices.

Detecting distracted driving accurately and quickly with limited resources is an essential yet underexplored problem. Most of the existing works ignore the resource-limited reality. In this work, we aim to achieve accurate and fast distracted driver detection in the context of embedded devices where only limited memory and computing resources are available. Specifically, we propose a novel convolutional neural network (CNN) light-weighting method via adjusting block layers and shrinking network channels without compromising the model's accuracy. Finally, the model is deployed on multiple devices with real-time detection of driving behaviour. The experimental results for the American University in Cairo (AUC) and StateFarm datasets demonstrate the effectiveness of the proposed method. For instance, for the AUC dataset, the proposed MobileNetV2-tiny model achieves 1.63% higher accuracy with just 78% of the model parameters of the original MobileNetV2 model. The inference speed of the proposed MobileNetV2-tiny model on resource-limited devices is on average 1.5 times that of the original MobileNetV2 model, which can meet real-time requirements.

Asymmetrical alterations of grey matter among psychiatric disorders: A systematic analysis by voxel-based activation likelihood estimation.

Schizophrenia (SZ), bipolar disorder (BD) and major depression disorder (MDD) have been regarded as highly diverged independent entities in current psychiatric diagnosis. However, ample new evidence suggests that they may have common biological traits. Neuroimaging studies showed that psychiatric disorders might associated with altered grey matter (GM) asymmetry compared to controls; however, the degree to which SZ, BD and MDD have common and/or distinct asymmetrical alterations in GM is still ambiguous. In this study, we analysed 169 voxel-based studies (including 3517 SZ patients, 1575 BD patients, 3280 MDD patients and 9733 controls) using activation likelihood estimation (ALE) meta-analysis to systematically review the existence of similar GM atrophy and asymmetrical alteration patterns among these psychiatric disorders, and the functional association between behaviour domains and topological alterations. We found that the right parahippocampal gyrus and left superior frontal gyrus showed commonly altered GM volume across all three illnesses, but did not identify common asymmetrical alteration. The asymmetrical alteration with leftward bias appeared in SZ and bipolar disorder at different locations, but more asymmetrical alteration with rightward bias appeared in MDD. Moreover, these changes have been confirmed to be associate with several symptoms and may have roles in functional networks. Our findings support the existence of common neurobiological damnification in these psychiatric disorders and provides valuable insights for the neural commonalties among different psychiatric disorders based on a large sample size.

A Deep Learning Framework for Driving Behavior Identification on In-Vehicle CAN-BUS Sensor Data.

Human driving behaviors are personalized and unique, and the automobile fingerprint of drivers could be helpful to automatically identify different driving behaviors and further be applied in fields such as auto-theft systems. Current research suggests that in-vehicle Controller Area Network-BUS (CAN-BUS) data can be used as an effective representation of driving behavior for recognizing different drivers. However, it is difficult to capture complex temporal features of driving behaviors in traditional methods. This paper proposes an end-to-end deep learning framework by fusing convolutional neural networks and recurrent neural networks with an attention mechanism, which is more suitable for time series CAN-BUS sensor data. The proposed method can automatically learn features of driving behaviors and model temporal features without professional knowledge in features modeling. Moreover, the method can capture salient structure features of high-dimensional sensor data and explore the correlations among multi-sensor data for rich feature representations of driving behaviors. Experimental results show that the proposed framework performs well in the real world driving behavior identification task, outperforming the state-of-the-art methods.

Characterization of fine motor development: dynamic analysis of children's drawing movements.

In this study, we investigated children's fine motor development by analyzing drawing trajectories, kinematics and kinetics. Straight lines drawing task and circles drawing task were performed by using a force sensitive tablet. Forty right-handed and Chinese mother-tongue students aged 6-12, attending classes from grade 1 to 5, were engaged in the experiment. Three spatial parameters, namely cumulative trace length, vector length of straight line and vertical diameter of circle were determined. Drawing duration, mean drawing velocity, and number of peaks in stroke velocity profile (NPV) were derived as kinematic parameters. Besides mean normal force, two kinetic indices were proposed: normalized force angle regulation (NFR) and variation of fine motor control (VFC) for circles drawing task. The maturation and automation of fine motor ability were reflected by increased drawing velocity, reduced drawing duration, NPV and NFR, with decreased VFC in circles drawing task. Grade and task main effects as well as significant correlations between age and parameters suggest that factors such as schooling, age and task should be considered in the assessment of fine motor skills. Compared with kinematic parameters, findings of NFR and VFC revealed that kinetics is another important perspective in the analysis of fine motor movement.