Research on Athlete Behavior Recognition Technology in Sports Teaching Video Based on Deep Neural Network

In recent years, due to the simple design idea and good recognition effect, deep learning method has attracted more and more researchers’ attention in computer vision tasks. Aiming at the problem of athlete behavior recognition in mass sports teaching video, this paper takes depth video as the research object and cuts the frame sequence as the input of depth neural network model, inspired by the successful application of depth neural network based on two-dimensional convolution in image detection and recognition. A depth neural network based on three-dimensional convolution is constructed to automatically learn the temporal and spatial characteristics of athletes’ behavior. The training results on UTKinect-Action3D and MSR-Action3D public datasets show that the algorithm can correctly detect athletes’ behaviors and actions and show stronger recognition ability to the algorithm compared with the images without clipping frames, which effectively improves the recognition effect of physical education teaching videos.

COVID-19 and thyroid disease: an infodemiological pilot study

Introduction: Some studies have linked COVID-19 with thyroid disease. Google Trends (GT) searches may reflect disease epidemiology. Recently, GT searches for COVID-19-associated terms have been linked to the epidemiology of COVID-19. In this study we aimed to assess COVID-19 cases per se vs COVID-19-associated GT searches and thyroid-associated GT searches. Materials-Methods: We collected data on worldwide weekly GT searches regarding “COVID-19”, “SARS-COV-2”, “coronavirus”, “smell”, “taste”, “fatigue”, “cough”, “thyroid”, “thyroiditis” and “subacute thyroiditis” for 92 weeks and worldwide weekly COVID-19 cases' statistics in the same time period. The study period was split in half and in each time period we performed cross-correlation analysis and mediation analysis. Results Significant positive CCF values were noted in both time periods; while COVID-19 cases per se were associated with “thyroid” searches in both time periods, significant CCFs for “fatigue”, “COVID-19” and “SARS-COV-s” were mostly found in the second time period. In the latter period, the effect of “COVID-19” searches on “thyroid” searches was significantly mediated by COVID-19 cases (p=0.048). Discussion COVID-19 cases per se were found to be associated with no lag with GT searches for COVID-19 symptoms in the first time period and in the second time period to lead searches for symptoms, COVID-19 terms as well as thyroid terms. Searches for a non-specific symptom or COVID-19 search terms mostly lead GT “thyroid” searches, in the second time period. This time frame/sequence particularly in the second time period (noted by the preponderance of the SARS-COV-2 delta variant), lends some credence to associations of COVID-19 cases per se with (apparent) thyroid disease (via searches for them).

Molecular characterization and immune protection of the 3-hydroxyacyl-CoA dehydrogenase gene in Echinococcus granulosus

Background Cystic echinococcosis (CE) is a serious parasitic zoonosis caused by the larvae of the tapeworm Echinococcus granulosus. The development of an effective vaccine is one of the most promising strategies for controlling CE. Methods The E. granulosus 3-hydroxyacyl-CoA dehydrogenase (EgHCDH) gene was cloned and expressed in Escherichia coli. The distribution of EgHCDH in protoscoleces (PSCs) and adult worms was analyzed using immunofluorescence. The transcript levels of EgHCDH in PSCs and adult worms were analyzed using quantitative real-time reverse transcription PCR (RT-qPCR). The immune protective effects of the rEgHCDH were evaluated. Results The 924-bp open reading frame sequence of EgHCDH, which encodes a protein of approximately 34 kDa, was obtained. RT-qPCR analysis revealed that EgHCDH was expressed in both the PSCs and adult worms of E. granulosus. Immunofluorescence analysis showed that EgHCDH was mainly localized in the tegument of PSCs and adult worms. Western blot analysis showed that the recombinant protein was recognized by E. granulosus-infected dog sera. Animal challenge experiments demonstrated that dogs immunized with recombinant (r)EgHCDH had significantly higher serum IgG, interferon gamma and interleukin-4 concentrations than the phosphate-buffered saline (PBS) control group. The rEgHCDH vaccine was able to significantly reduce the number of E. granulosus and inhibit the segmental development of E. granulosus compared to the PBS control group. Conclusions The results suggest that rEgHCDH can induce partial immune protection against infection with E. granulosus and could be an effective candidate for the development of new vaccines. Graphical abstract

CSCD2: an integrated interactional database of cancer-specific circular RNAs

The significant function of circRNAs in cancer was recognized in recent work, so a well-organized resource is required for characterizing the interactions between circRNAs and other functional molecules (such as microRNA and RNA-binding protein) in cancer. We previously developed cancer-specific circRNA database (CSCD), a comprehensive database for cancer-specific circRNAs, which is widely used in circRNA research. Here, we updated CSCD to CSCD2 (http://geneyun.net/CSCD2 or http://gb.whu.edu.cn/CSCD2), which includes significantly more cancer-specific circRNAs identified from a large number of human cancer and normal tissues/cell lines. CSCD2 contains >1000 samples (825 tissues and 288 cell lines) and identifies a large number of circRNAs: 1 013 461 cancer-specific circRNAs, 1 533 704 circRNAs from only normal samples and 354 422 circRNAs from both cancer and normal samples. In addition, CSCD2 predicts potential miRNA–circRNA and RBP–circRNA interactions using binding motifs from >200 RBPs and 2000 microRNAs. Furthermore, the potential full-length and open reading frame sequence of these circRNAs were also predicted. Collectively, CSCD2 provides a significantly enhanced resource for exploring the function and regulation of circRNAs in cancer.

Channel-Wise Spatiotemporal Aggregation Technology for Face Video Forensics

Recent progress in deep learning, in particular the generative models, makes it easier to synthesize sophisticated forged faces in videos, leading to severe threats on social media about personal privacy and reputation. It is therefore highly necessary to develop forensics approaches to distinguish those forged videos from the authentic. Existing works are absorbed in exploring frame-level cues but insufficient in leveraging affluent temporal information. Although some approaches identify forgeries from the perspective of motion inconsistency, there is so far not a promising spatiotemporal feature fusion strategy. Towards this end, we propose the Channel-Wise Spatiotemporal Aggregation (CWSA) module to fuse deep features of continuous video frames without any recurrent units. Our approach starts by cropping the face region with some background remained, which transforms the learning objective from manipulations to the difference between pristine and manipulated pixels. A deep convolutional neural network (CNN) with skip connections that are conducive to the preservation of detection-helpful low-level features is then utilized to extract frame-level features. The CWSA module finally makes the real or fake decision by aggregating deep features of the frame sequence. Evaluation against a list of large facial video manipulation benchmarks has illustrated its effectiveness. On all three datasets, FaceForensics++, Celeb-DF, and DeepFake Detection Challenge Preview, the proposed approach outperforms the state-of-the-art methods with significant advantages.

Video Based Suspicious Human Behaviour Recognition System

In recent years, the number of surveillance cameras installed to monitor private and public spaces has increased rapidly. The demand has raised for smarter video surveillance of public and private spaces using intelligent vision systems which can differentiate between 'suspicious' and 'unsuspicious' behaviours according to the human observer. Generally, the video streams are constantly recorded or monitored by operators. In these cases, an intelligent system can give more accurate performance than a human. We have proposed a method called motion influence map under machine learning for representing human activities. Optical-flow is computed for each pixel in a frame that are processed sequentially. The key feature of the proposed motion influence map is that it effectively reflects the motion characteristics such as movement speed, movement direction, and size of the objects or subjects and their interactions within a frame sequence. It further extracts frames of high motion influence values and compares with the testing frames to automatically detect suspicious activities.

Stochastic Neural Networks for Automatic Cell Tracking in Microscopy Image Sequences of Bacterial Colonies

We describe an automated analysis method to quantify the detailed growth dynamics of a population of bacilliform bacteria. We propose an innovative approach to frame-sequence tracking of deformable-cell motion by the automated minimization of a new, specific cost functional. This minimization is implemented by dedicated Boltzmann machines (stochastic recurrent neural networks). Automated detection of cell divisions is handled similarly by successive minimizations of two cost functions, alternating the identification of children pairs and parent identification. We validate this automatic cell tracking algorithm using recordings of simulated cell colonies that closely mimic the growth dynamics of E. coli in microfluidic traps. On a batch of 1100 image frames, cell registration accuracies per frame ranged from 94.5% to 100%, with a high average. Our initial tests using experimental image sequences of E. coli colonies also yield convincing results, with a registration accuracy ranging from 90% to 100%.