Search

We evaluated the relative attribution and interactions of treatment and patient-related risk factors for second primary malignancies (SPMs) in cervical and endometrial cancer survivors.Stage I-III cervical and endometrial cancer survivors' data from the Surveillance, Epidemiology, and End Results (SEER) registry between January 1988 and December 2015 were analyzed. The standardized incidence ratio (SIR), excess absolute risk (EAR), and corresponding 95% confidence interval (95% CI) values were calculated. Analyses were classified based on proxies of human papillomavirus (HPV), smoking, hormone, and radiotherapy (RT) status. Additive and multiplicative interactions were assessed.Cervical cancer survivors had a higher risk for developing potentially HPV and smoking-related SPMs, especially in the RT group (SIRRT, HPV, and smoking promote SPMs in cervical cancer to different extents. The SPM burden in cervical cancer survivors could be mostly attributed to smoking and RT and their interactions.

The ever-increasing computation complexity of fastgrowing Deep Neural Networks (DNNs) has requested new computing paradigms to overcome the memory wall in conventional Von Neumann computing architectures. The emerging Computing-In-Memory (CIM) architecture has been a promising candidate to accelerate neural network computing. However, data movement between CIM arrays may still dominate the total power consumption in conventional designs. This paper proposes a flexible CIM processor architecture named Domino and "Computing-On-the-Move" (COM) dataflow, to enable stream computing and local data access to significantly reduce data movement energy. Meanwhile, Domino employs customized distributed instruction scheduling within Network-on-Chip (NoC) to implement inter-memory computing and attain mapping flexibility. The evaluation with prevailing DNN models shows that Domino achieves 1.77-to-2.37$\times$ power efficiency over several state-of-the-art CIM accelerators and improves the throughput by 1.28-to-13.16$\times$., Comment: arXiv admin note: text overlap with arXiv:2107.09500

Background Nonalcoholic fatty liver disease (NAFLD) is a public health challenge and significant cause of morbidity and mortality worldwide. Early identification is crucial for disease intervention. We recently proposed a nomogram-based NAFLD prediction model from a large population cohort. We aimed to explore machine learning tools in predicting NAFLD. Methods A retrospective cross-sectional study was performed on 15 315 Chinese subjects (10 373 training and 4942 testing sets). Selected clinical and biochemical factors were evaluated by different types of machine learning algorithms to develop and validate seven predictive models. Nine evaluation indicators including area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), accuracy, positive predictive value, sensitivity, F1 score, Matthews correlation coefficient (MCC), specificity and negative prognostic value were applied to compare the performance among the models. The selected clinical and biochemical factors were ranked according to the importance in prediction ability. Results Totally 4018/10 373 (38.74%) and 1860/4942 (37.64%) subjects had ultrasound-proven NAFLD in the training and testing sets, respectively. Seven machine learning based models were developed and demonstrated good performance in predicting NAFLD. Among these models, the XGBoost model revealed the highest AUROC (0.873), AUPRC (0.810), accuracy (0.795), positive predictive value (0.806), F1 score (0.695), MCC (0.557), specificity (0.909), demonstrating the best prediction ability among the built models. Body mass index was the most valuable indicator to predict NAFLD according to the feature ranking scores. Conclusions The XGBoost model has the best overall prediction ability for diagnosing NAFLD. The novel machine learning tools provide considerable beneficial potential in NAFLD screening.

Image inpainting has achieved fundamental advances with deep learning. However, almost all existing inpainting methods aim to process natural images, while few target Thermal Infrared (TIR) images, which have widespread applications. When applied to TIR images, conventional inpainting methods usually generate distorted or blurry content. In this paper, we propose a novel task -- Thermal Infrared Image Inpainting, which aims to reconstruct missing regions of TIR images. Crucially, we propose a novel deep-learning-based model TIR-Fill. We adopt the edge generator to complete the canny edges of broken TIR images. The completed edges are projected to the normalization weights and biases to enhance edge awareness of the model. In addition, a refinement network based on gated convolution is employed to improve TIR image consistency. The experiments demonstrate that our method outperforms state-of-the-art image inpainting approaches on FLIR thermal dataset.

The data storage has been one of the bottlenecks in surveillance systems. The conventional video compression schemes such as H.264 and H.265 do not fully utilize the low information density characteristic of the surveillance video, and they attach equal importance to foreground and background when performing compression. In this article, we propose a novel video compression scheme that compresses the foreground and background of the surveillance video separately. The compression ratio is greatly improved by sharing background information among adjacent frames through an adaptive background updating and interpolation module. Besides, we present two different schemes to compress the foreground and compare their performance in the ablation study to show the importance of temporal information for video compression. In the decoding end, a coarse-to-fine two-stage module is applied to achieve the composition of the foreground and background and the enhancements of frame quality. The experimental results show that our proposed method requires 49.75% less bpp (bits per pixel) than the conventional algorithm H.265 to achieve the same PSNR (36 dB) on the HEVC dataset.

Automated medical image segmentation for organs or lesions plays an essential role in clinical diagnoses and treatment plannings. However, training an accurate and robust segmentation model is still a long-standing challenge due to the time-consuming and expertise-intensive annotations for training data, especially 3-D medical images. Recently, self-supervised learning emerges as a promising approach for unsupervised visual representation learning, showing great potential to alleviate the expertise annotations for medical images. Although global representation learning has attained remarkable results on iconic datasets, such as ImageNet, it can not be applied directly to medical image segmentation, because the segmentation task is non-iconic, and the targets always vary in physical scales. To address these problems, we propose a Multi-scale Visual Representation self-supervised Learning (MsVRL) model, to perform finer-grained representation and deal with different target scales. Specifically, a multi-scale representation conception, a canvas matching method, an embedding pre-sampling module, a center-ness branch, and a cross-level consistent loss are introduced to improve the performance. After pre-trained on unlabeled datasets (RibFrac and part of MSD), MsVRL performs downstream segmentation tasks on labeled datasets (BCV, spleen of MSD, and KiTS). Results of the experiments show that MsVRL outperforms other state-of-the-art works on these medical image segmentation tasks.

Background The survival advantage of radiotherapy for patients with stage IV classic Hodgkin lymphoma (HL) has not been adequately evaluated. Methods We analyzed patients with stage IV HL enrolled from the Surveillance, Epidemiology, and End Results (SEER) registry from January 2000 to December 2012. Propensity score (PS) analysis with 1:2 matching was performed to ensure well-balanced characteristics of the comparison groups. Kaplan-Meier and Cox proportional hazardous model were used to evaluate the overall survival (OS), cancer-specific survival (CSS), the hazards ratio (HR) and corresponding 95% confidence intervals (95% CI). Results Overall, for all patients with stage IV HL, receiving radiotherapy was associated with both significantly improved OS and CSS. Radiotherapy to any lesions could independently improve the OS and CSS by 30% to 36% in the multivariate analyses before and after PS matching (PSM), with the best improvement of 33% to 40% observed for patients with nodular sclerosis (P < 0.05) among all HL pathological types. In particular, radiotherapy, most likely to the residual site, was more pronouncedly associated with the improvement in survival for patients with stage IV HL who were young (age

The Artificial Intelligence (AI) in edge computing is requesting new processing units with a much higher computing-power ratio. The emerging resistive Non-Volatile Memory (NVM) with the in-memory computing capability may greatly advance the AI hardware technologies. In this paper, we propose the use of binary resistive memory to form an 8-bit fixed-point data/weight for AI computing. A robust Computing-In-Memory (CIM) core with digital input and analog output Multiplication-and-Accumulation (MAC) circuit is proposed. The corresponding integration scheme and Successive Approximation Register Analog-to-Digital Converter (SAR ADC) based data conversion scheme are also presented. The simulation results show that the proposed CIM core achieves 7.26 bit of Effective Number of Bits (ENOB) with 0.78mW (256*1) power consumption and 1.85M/s computing speed. Compared with previously reported CIM implementations and Deep Learning Accelerators (DLAs) (without CIM ability), our design achieves 2.23– $7.26\times $ better energy efficiency in 8-bit input 8-bit weight pattern, and achieves relatively high accuracy with LeNet and AlexNet.

Learning accurate low-dimensional embeddings for a network is a crucial task as it facilitates many network analytics tasks. Moreover, the trained embeddings often require a significant amount of space to store, making storage and processing a challenge, especially as large-scale networks become more prevalent. In this paper, we present a novel semi-supervised network embedding and compression method, SNEQ, that is competitive with state-of-art embedding methods while being far more space- and time-efficient. SNEQ incorporates a novel quantisation method based on a self-attention layer that is trained in an end-to-end fashion, which is able to dramatically compress the size of the trained embeddings, thus reduces storage footprint and accelerates retrieval speed. Our evaluation on four real-world networks of diverse characteristics shows that SNEQ outperforms a number of state-of-the-art embedding methods in link prediction, node classification and node recommendation. Moreover, the quantised embedding shows a great advantage in terms of storage and time compared with continuous embeddings as well as hashing methods.

Background: A consensus regarding optimum treatment strategies for locally advanced gastric cancer (LAGC) has not yet been reached. We aimed to evaluate the efficacy of various treatment modalities for LAGC and provided clinicians salvage options under real-world situation. Methods: Medical charts of patients with LAGC who underwent radical resection plus adjuvant chemotherapy or chemoradiotherapy from July 2003 to December 2014 were included. Validation cohort were selected from SEER database between 2004 and 2014. Kaplan-Meier and Cox proportional hazardous models were used to evaluate the overall survival (OS), cancer-specific survival (CSS), and disease-free survival (DFS). Propensity score matching (PSM) was used to adjust for potential baseline confounding. Results: A total of 350 patients were included and divided into D1 dissection plus chemotherapy group (D1CT, n = 74), D1 dissection plus adjuvant chemoradiotherapy group (D1CRT, n = 69), D2 dissection plus adjuvant chemotherapy group (D2CT, n = 134), and D2 dissection plus adjuvant chemoradiotherapy group (D2CRT, n = 73). PSM identified 50 patients in each group. After PSM, better DFS (P for D2CRT vs. D1CT, D1CRT, and D2CT was 0.001, 0.006, and 0.001, respectively) and OS (P for D2CRT vs. D1CT, D1CRT, and D2CT was 0.001, 0.011, and 0.022, respectively) were found for the D2CRT group (mean, OS = 110.7months, DFS = 95.2 months) than the other groups. Similar findings were further validated in the Surveillance, Epidemiology, and End Results database (SEER) cohort. In addition, patients in the D1CRT group achieved similar survival outcomes to those in the D2CT group (mean OS, 72.8 vs. 59.1 months, P = 0.86; mean DFS, 54.4 vs. 34.1 months, P = 0.460). Conclusions: The results of the study indicated the better role for D2CRT in treating the LAGC, meanwhile, the patients treated with D1CRT might achieve similar survival as that of D2CT patients.

Electroencephalography (EEG) emotion recognition, an important task in Human-Computer Interaction (HCI), has made a great breakthrough with the help of deep learning algorithms. Although the application of attention mechanism on conventional models has improved its performance, most previous research rarely focused on multiplex EEG features jointly, lacking a compact model with unified attention modules. This study proposes Joint-Dimension-Aware Transformer (JDAT), a robust model based on squeezed Multi-head Self-Attention (MSA) mechanism for EEG emotion recognition. The adaptive squeezed MSA applied on multidimensional features enables JDAT to focus on diverse EEG information, including space, frequency, and time. Under the joint attention, JDAT is sensitive to the complicated brain activities, such as signal activation, phase-intensity couplings, and resonance. Moreover, its gradually compressed structure contains no recurrent or parallel modules, greatly reducing the memory and complexity, and accelerating the inference process. The proposed JDAT is evaluated on DEAP, DREAMER, and SEED datasets, and experimental results show that it outperforms state-of-the-art methods along with stronger flexibility.

• The level of NLR on admission could be an independent risk factor for the severe disease and the mortality of COVID-19. • The predictive value of NLR for poor prognosis was more significant in patients without other potential risk factors. • NLR could help physicians rapidly identify high-risk patients and adopt timely intervention.

To enhance the security level of digital information, the biometric authentication method based on Electrocardiographic (ECG) is gaining increasing attention in a wide range of applications. Compared with other biometric features, e.g., fingerprint and face, the ECG signals have several advantages, such as higher security, simpler acquisition, liveness detection, and health information. Therefore, various methods for ECG-based authentication have been proposed. However, the generalization ability of these methods is limited because the feature extraction for the ECG signals in conventional methods is data dependent. To improve the generalization ability and achieve more stable results on different datasets, a parallel multi-scale one-dimensional residual network is proposed in this paper. This network utilizes three convolutional kernels with different kernel sizes, achieving better classification accuracy than the conventional schemes. Moreover, two loss functions named center loss and margin loss are used during the training of the network. Compared with the conventional softmax loss, these two loss functions can further improve the generalization ability of the extracted embedding features. Furthermore, we evaluate the effectiveness of our proposed method thoroughly on the ECG-ID database, the PTB Diagnostic ECG database, and the MIT-BIH Arrhythmia database, achieving 2.00%, 0.59%, and 4.74% of equal error rate (EER), respectively. Compared with other works, our proposed method improves 1.61% and 4.89% classification accuracy on the ECG-ID database and the MIT-BIH Arrhythmia database, respectively.

The separation of the data capture and analysis in modern vision systems has led to a massive amount of data transfer between the end devices and cloud computers, resulting in long latency, slow response, and high power consumption. Efficient hardware architectures are under focused development to enable Artificial Intelligence (AI) at the resource-limited end sensing devices. One of the most promising solutions is to enable Processing-in-Pixel (PIP) scheme. However, the conventional schemes suffer from the low fill-factor issue. This paper proposes a PIP based CMOS sensor architecture, which allows convolution operation before the column readout circuit to significantly improve the image reading speed with much lower power consumption. The simulation results show that the proposed architecture could support the computing efficiency up to 11.65 TOPS/W at the 8-bit weight configuration, which is three times as high as the conventional schemes. The transistors required for each pixel are only 2.5T, significantly improving the fill-factor.

Human pose transfer aims at transferring the appearance of the source person to the target pose. Existing methods utilizing flow-based warping for non-rigid human image generation have achieved great success. However, they fail to preserve the appearance details in synthesized images since the spatial correlation between the source and target is not fully exploited. To this end, we propose the Flow-based Dual Attention GAN (FDA-GAN) to apply occlusion- and deformation-aware feature fusion for higher generation quality. Specifically, deformable local attention and flow similarity attention, constituting the dual attention mechanism, can derive the output features responsible for deformable- and occlusion-aware fusion, respectively. Besides, to maintain the pose and global position consistency in transferring, we design a pose normalization network for learning adaptive normalization from the target pose to the source person. Both qualitative and quantitative results show that our method outperforms state-of-the-art models in public iPER and DeepFashion datasets.

Traveling by bike-sharing systems has become an indispensable means of transportation in our daily lives because green commuting has gradually become a consensus and conscious action. However, the problem of “difficult to rent or to return a bike” has gradually become an issue in operating the bike-sharing system. Moreover, scientific and systematic schemes that can efficiently complete the task of rebalancing bike-sharing systems are lacking. This study aims to introduce the basic idea of the k-divisive hierarchical clustering algorithm. A rebalancing strategy based on the model of level of detail in combination with genetic algorithm was proposed. Data were collected from the bike-sharing system in Ningbo. Results showed that the proposed algorithm could alleviate the problem of the uneven distribution of the demand for renting or returning bikes and effectively improve the service from the bike-sharing system. Compared with the traditional method, this algorithm helps reduce the effective time for rebalancing bike-sharing systems by 28.3%. Therefore, it is an effective rebalancing scheme.

The morphology of Fetal Electrocardiogram (FECG) plays an important role in the early diagnosis of fetal health condition. However, it is intractable to extract the clean morphology of FECG signals, which are usually contaminated by Maternal ECG (MECG) and various noises. To extract the clean morphology of FECG signals from noninvasive abdominal ECG records, a high-performance and high-efficient two-stage Slow-Fast Long Short Term Memory (SFLSTM) based architecture is proposed. The MECG elimination and the FECG enhancement are realized by the elaborately designed slow LSTM and fast LSTM to filter out the MECG and the residual noise components, respectively. Qualitative and quantitative experiments are conducted on the records from two public databases. The experimental results show that our proposed MECG elimination and FECG enhancement schemes improve the Signal-to-Noise Ratio (SNR) by 3.09 dB and 1.81 dB, respectively. The proposed fast LSTM reduces the amountof computation by approximately 50%, without any degradation in performance. Our proposed method may leverage the noninvasive FECG monitoring for the early detection of fetal heart diseases.

To characterize the chest computed tomography (CT) findings of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) according to clinical severity. We compared the CT features of common cases and severe cases, symptomatic patients and asymptomatic patients, and febrile and afebrile patients. This was a retrospective analysis of the clinical and thoracic CT features of 120 consecutive patients with confirmed SARS-CoV-2 pneumonia admitted to a tertiary university hospital between January 10 and February 10, 2020, in Wuhan city, China. On admission, the patients generally complained of fever, cough, shortness of breath, and myalgia or fatigue, with diarrhea often present in severe cases. Severe patients were 20 years older on average and had comorbidities and an elevated lactate dehydrogenase (LDH) level. There were no differences in the CT findings between asymptomatic and symptomatic common type patients or between afebrile and febrile patients, defined according to Chinese National Health Commission guidelines. The clinical and CT features at admission may enable clinicians to promptly evaluate the prognosis of patients with SARS-CoV-2 pneumonia. Clinicians should be aware that clinically silent cases may present with CT features similar to those of symptomatic common patients. • The clinical features and predominant patterns of abnormalities on CT for asymptomatic, typic common, and severe cases were summarized. These findings may help clinicians to identify severe patients quickly at admission. • Clinicians should be cautious that CT findings of afebrile/asymptomatic patients are not better than the findings of other types of patients. These patients should also be quarantined. • The use of chest CT as the main screening method in epidemic areas is recommended.

The method of importance map has been widely adopted in DNN-based lossy image compression to achieve bit allocation according to the importance of image contents. However, insufficient allocation of bits in non-important regions often leads to severe distortion at low bpp (bits per pixel), which hampers the development of efficient content-weighted image compression systems. This paper rethinks content-based compression by using Generative Adversarial Network (GAN) to reconstruct the non-important regions. Moreover, multiscale pyramid decomposition is applied to both the encoder and the discriminator to achieve global compression of high-resolution images. A tunable compression scheme is also proposed in this paper to compress an image to any specific compression ratio without retraining the model. The experimental results show that our proposed method improves MS-SSIM by more than 10.3% compared to the recently reported GAN-based method [3] to achieve the same low bpp (0.05) on the Kodak dataset.

Deep convolutional neural networks (CNNs) have achieved state-of-the-art performance in a wide range of applications. However, deeper CNN models, which are usually computation consuming, are widely required for complex artificial intelligence (AI) tasks. Though recent research progress on network compression, such as pruning, has emerged as a promising direction to mitigate computational burden, existing accelerators are still prevented from completely utilizing the benefits of leveraging sparsity due to the irregularity caused by pruning. On the other hand, field-programmable gate arrays (FPGAs) have been regarded as a promising hardware platform for CNN inference acceleration. However, most existing FPGA accelerators focus on dense CNN and cannot address the irregularity problem. In this article, we propose a sparsewise dataflow to skip the cycles of processing multiply-and-accumulates (MACs) with zero weights and exploit data statistics to minimize energy through zeros gating to avoid unnecessary computations. The proposed sparsewise dataflow leads to a low bandwidth requirement and high data sharing. Then, we design an FPGA accelerator containing a vector generator module (VGM) that can match the index between sparse weights and input activations according to the proposed dataflow. Experimental results demonstrate that our implementation can achieve 987-, 46-, and 57-imag/s performance for AlexNet, VGG-16, and ResNet-50 on Xilinx ZCU102, respectively, which provides $1.5\times $ – $6.7\times $ speedup and $2.0\times $ – $6.0\times $ energy efficiency over previous CNN FPGA accelerators.

An efficient method is proposed for predicting the attenuation and distortion of the damped oscillation surges traveling on single-phase overhead lines under the influence of corona and frequency dependence of line parameters. The finite-difference time-domain algorithm is employed in the transient computation. The Suliciu's corona model is extended to simulate the charge–voltage curves with minor loops under the damped oscillation surges. The corona effect is considered by adding a corona current term in the telegraphers' equations. The frequency-dependence effect is also taken into account by using the vector fitting and recursive convolution in the time domain. The computed voltage waveforms are compared with the measured ones to examine the validity of the proposed method. The computed results show that the minor loops have a greater effect on the attenuation and distortion for the damped oscillation surges under negative polarity.

The large area and high power consumption are the two main bottlenecks in the conventional SRAM-based Field Programmable Gate Arrays (FPGAs). In recent works, resistive Non-Volatile Memories (NVMs) have been widely proposed to tackle the above issues in the reconfigurable computing systems, due to their non-volatility, fast read/write speed and high-density. The magnetic Domain-Wall (DW) Racetrack Memory (RM) is the emerging NVM with the great prospect of the development of the low-power and high-density circuits and systems. This paper presents RM based single-context and multi-context hybrid Look-Up Tables (LUTs). The hybrid structure allows the LUT to support both volatile input (low-power and high-speed input) and non-volatile input. The non-volatile input is used to reduce the leakage power and also to provide additional reusable resources to increase the hardware utilization. Compared to the SRAM-based 6-input LUT, the proposed non-volatile LUT reduces the number of transistors and leakage power by 80.2% and 84.2%, respectively. The proposed design also reduces the leakage power of the conventional 6-input non-volatile LUT by 17.4% with 27.3% fewer transistors and 36% faster operation speed. The Verilog-to-Routing (VTR) simulation results show that the proposed 6-input LUT consumes 27.1% less power than the SRAM-based counterpart. It may also provide 15.2% additional reusable resource. The context of the proposed multi-context LUT can be switched in 4 ns with the context switching energy of 397.24 fJ/LUT.

The influence of corona on the propagation of transient voltages on transmission lines can be studied by modeling the impulse corona discharge, taking into account its physical properties. The technique of corona characteristics (charge-voltage) is more favorable than the propagation delay characteristic which is known to be less accurate and leads to results that are far from reality. In this paper, a mathematical model of corona discharge is integrated into the Alternative Transients Program of the ElectroMagnetic Transients Program software (ATP/EMTP), using the MODELS interface. The attenuation and distortion phenomena on the overhead transmission line caused by corona is modeled and simulated by the implementation of a corona model with the user-defined multi-branch type-94 circuit in ATP. Comparison between the computed results and literature experimental investigations gives a good agreement between them.

A high selectivity semiconductor selector is the key component in the ultra-high density and low power 3-D resistive non-volatile memory array. The state-of-art selectors suffer from various issues including low selectivity, high OFF current, and low ON current, which significantly limit the array size and performance. Recently, a hysteresis loop (HL) selector with the high selectivity, moderate ON voltage and large HL window was developed to address the high leakage current issue. In this brief, a two-step sensing scheme is proposed to minimize the read disturbance and sensing power. The proposed two-step sensing circuit could achieve 25-ns fast sensing speed with 3.92-uW low sensing power.

The data movement between the processing and storage units has been one of the most critical issues in modern computer systems. The emerging Resistive Random Access Memory (RRAM) technology has drawn tremendous attention due to its non-volatile ability and the potential in computation application. These properties make them a perfect choice for application in modern computing systems. In this paper, an 8-bit radix-4 non-volatile parallel multiplier is proposed, with improved computational capabilities. The corresponding booth encoding scheme, read-out circuit, simplified Wallace tree, and Manchester carry chain are presented, which help to short the delay of the proposed multiplier. While the presence of RRAM save computational time and overall power as multiplicand is stored beforehand. The area of the proposed non-volatile multiplier is reduced with improved computing speed. The proposed multiplier has an area of 785.2 μm2 with Generic Processing Design Kit 45 nm process. The simulation results show that the proposed multiplier structure has a low computing power at 161.19 μW and a short delay of 0.83 ns with 1.2 V supply voltage. Comparative analyses are performed to demonstrate the effectiveness of the proposed multiplier design. Compared with conventional booth multipliers, the proposed multiplier structure reduces the energy and delay by more than 70% and 19%, respectively.

Fetal electrocardiogram (FECG) morphology plays an essential role in the early diagnosis of fetal health conditions. However, it is intractable to extract the clean morphology of FECG signals, which are usually contaminated by maternal ECG (MECG) and various noises. To extract the clean morphology of FECG signals from non-invasive abdominal ECG records, a high-performance and high-efficient two-stage slow-fast long short-term memory (SFLSTM) based architecture is proposed. The MECG elimination and the FECG enhancement are realized by the elaborately designed slow LSTM and fast LSTM to filter out the MECG and the residual noise components. Qualitative and quantitative experiments are conducted on the records from two public datasets. The experimental results reveal that our designed scheme achieves the best performance in k SQI , signal-to-noise ratio (SNR), and root mean square error (RMSE). The MECG elimination and the FECG enhancement improve SNR by 3.09 and 1.81 dB, respectively. The proposed fast LSTM reduces computation cost by approximately 50%, without any degradation in performance. Our method may leverage non-invasive FECG monitoring for the early detection of fetal heart diseases.

Iris centre is widely used in feature-based gaze estimation system. However, the performance of various existing iris centre localisation methods is degraded in low-resolution images. To address the problem, an accurate regression-based gaze estimation system is proposed by replacing the iris centre with a novel average-binary-connected-component-centroid (ABCCC) feature. First, ABCCC is obtained by averaging a series of binary connected component centroids, which are calculated by different percentile grey thresholds in the normalised eye region of interest. Then, head pose is estimated by employing the OpenCV iterative algorithm. Thereafter, ABCCC and head pose are mapped to compute the gaze using a linear mapping functions, which are regressed in calibration phase. Compared with traditional methods using iris centre, ABCCC is easier to calculate and shows a stronger correlation with the change of gaze, which improves the accuracy of the gaze estimation system. Experimental results on EYEDIAP database verify the effectiveness of the proposed method.

The survival advantage of surgery to the primary tumor for patients with distant metastatic esophageal cancer has not been adequately evaluated. This study aims to investigate the role of surgery to the primary tumor in distant metastatic esophageal cancer and to evaluate possible different effects of surgery on survival of esophageal adenocarcinoma (EAC) and esophageal squamous cell carcinoma (ESCC). This study included a cohort of 4,367 metastatic esophageal cancer patients from the Surveillance, Epidemiology, and End Results (SEER) database, registered from January 2004 to December 2014. Kaplan-Meier and Cox proportional hazardous models were used to evaluate the overall survival (OS) and corresponding 95% confidence interval (CI). Propensity score matching (PSM) was used to adjust for potential baseline confounding. Both EAC (median OS for surgery group vs. no-surgery group-14.0 vs. 9.0 months, P < 0.001) and ESCC (median OS for surgery vs. no-surgery group-11.0 vs. 7.0 months, P = 0.002) experienced survival benefits from surgery. We found that surgery to the primary tumor, when combined with chemotherapy, was associated with improved survival for patients with M1b disease, both EAC and ESCC, with a greater benefit observed in younger patients, and those with EAC. While the present data indicate a potential survival benefit from surgery for some patients with metastatic esophageal cancer, it is possible that performance status and metastatic disease burden impacted patient selection, influencing these results. Further studies are needed to determine the role of surgery for patients with metastatic esophageal cancer.

Human video motion transfer has a wide range of applications in multimedia, computer vision and graphics. Recently, due to the rapid development of Generative Adversarial Networks (GANs), there has been significant progress in the field. However, almost all existing GAN-based works are prone to address the mapping from human motions to video scenes, with scene appearances are encoded individually in the trained models. Therefore, each trained model can only generate videos with a specific scene appearance, new models are required to be trained to generate new appearances. Besides, existing works lack the capability of appearance control. For example, users have to provide video records of wearing new clothes or performing in new backgrounds to enable clothes or background changing in their synthetic videos, which greatly limits the application flexibility. In this paper, we propose GAC-GAN, a general method for appearance-controllable human video motion transfer. To enable general-purpose appearance synthesis, we propose to include appearance information in the conditioning inputs. Thus, once trained, our model can generate new appearances by altering the input appearance information. To achieve appearance control, we first obtain the appearance-controllable conditioning inputs and then utilize a two-stage GAC-GAN to generate the corresponding appearance-controllable outputs, where we utilize an ACGAN loss and a shadow extraction module for output foreground and background appearance control respectively. We further build a solo dance dataset containing a large number of dance videos for training and evaluation. Experimental results show that, our proposed GAC-GAN can not only support appearance-controllable human video motion transfer but also achieve higher video quality than state-of-art methods., Comment: paper under review

On the basis of the microphysical processes of gas discharge, a mathematical physical model of the corona under damped oscillation impulses is proposed in this study. The effects of collision ionisation, drift of the charged particles, space photoionisation, attachment and recombination are all taken into account in the proposed model. The space photoionisation formula is deduced in a coaxial cylindrical geometry. The charge–voltage characteristics under damped oscillation and double exponential impulses can be computed by the proposed model. The computed results are compared with the measured ones and a better agreement is shown between them. The main difference on the charge–voltage characteristics under the two types of impulses is also investigated by the model computation and experimental measurement.

Background: The stage-specific roles of radiotherapy (RT) alone, chemotherapy alone, and chemoradiotherapy (CRT) for patients with nodular lymphocyte-predominant Hodgkin lymphoma (NLPHL) have not been adequately evaluated. Methods: We analyzed patients with NLPHL in all stages enrolled from the Surveillance, Epidemiology, and End Results (SEER) registry from January 2000 to December 2015. Propensity score (PS) analysis with 1:1 matching was performed to ensure well-balanced characteristics between the comparison groups. Kaplan-Meier and Cox proportional hazards models were used to evaluate the overall survival (OS), cancer-specific survival (CSS), the hazard ratio (HR), and corresponding 95% confidence intervals (95% CI). Restricted mean survival times (RMST) were also used for the survival analyses. Results: For early-stage patients, RT alone was found to be associated with the improved survival of 10 months to 15 months and the deceased HR of 60% to 80%. Chemotherapy alone also had survival benefits (improved OS = 12.1 months; improved CSS = 5.3 months). However, not all of these improvements were statistically significant. CRT was associated with the best survival: the mean OS was significantly improved by 20 months and the HR was reduced by more than 80%, both before and after PS matching (PSM) (P < 0.05). For advanced-stage patients, none of RT alone, chemotherapy alone, and CRT had a significant effect on survival. Chemotherapy alone and CRT might be more beneficial for long-term survival (RMST120m: neither RT nor chemotherapy vs. chemotherapy alone vs. CRT = 104m vs. 111m vs. 108m). Subgroup analysis showed that the efficacy of RT alone in patients with early-stage NLPHL was significantly different between younger people and older people (age < 40: HR, 95% CI = 5.70, 0.34 - 94.2, P = 0.224; age ≧ 40: HR, 95% CI = 0.06, 0.01 - 0.32, P = 0.001; P for interaction = 0.005). Conclusions: The results from the large SEER database suggested that CRT was associated with the best survival of patients with the early-stage NLPHL. More effective treatment strategies for patients with advanced-stage NLPHL remain to be further studied. Citation Format: Shijie Wang, Mingfang Jia, Jianglong Han, Rui Zhang, Kejie Huang, Yunfeng Qiao, Ping Chen, Zhenming Fu. The stage-specific roles of radiotherapy and chemotherapy in nodular lymphocyte-predominant Hodgkin lymphoma patients: A propensity-matched analysis of the Surveillance, Epidemiology, and End Results database [abstract]. In: Proceedings of the AACR Virtual Meeting: Advances in Malignant Lymphoma; 2020 Aug 17-19. Philadelphia (PA): AACR; Blood Cancer Discov 2020;1(3_Suppl):Abstract nr PO-60.