A hybrid approach for categorizing images based on complex networks and neural networks

There are several methods for categorizing images, the most of which are statistical, geometric, model-based and structural methods. In this paper, a new method for describing images based on complex network models is presented. Each image contains a number of key points that can be identified through standard edge detection algorithms. To understand each image better, we can use these points to create a graph of the image. In order to facilitate the use of graphs, generated graphs are created in the form of a complex network of small-worlds. Complex grid features such as topological and dynamic features can be used to display image-related features. After generating this information, it normalizes them and uses them as suitable features for categorizing images. For this purpose, the generated information is given to the neural network. Based on these features and the use of neural networks, comparisons between new images are performed. The results of the article show that this method has a good performance in identifying similarities and finally categorizing them.

Advanced eNose-Driven Pedestrian Tracking Pipeline for Intelligent Car Driver Assisting System: Preliminary Results

From a biological point of view, alcohol human attentional impairment occurs before reaching a Blood Alcohol Content (BAC index) of 0.08% (0.05% under the Italian legislation), thus generating a significant impact on driving safety if the drinker subject is driving a car. Car drivers must keep a safe driving dynamic, having an unaltered physiological status while processing the surrounding information coming from the driving scenario (e.g., traffic signs, other vehicles and pedestrians). Specifically, the identification and tracking of pedestrians in the driving scene is a widely investigated problem in the scientific community. The authors propose a full, deep pipeline for the identification, monitoring and tracking of the salient pedestrians, combined with an intelligent electronic alcohol sensing system to properly assess the physiological status of the driver. More in detail, the authors propose an intelligent sensing system that makes a common air quality sensor selective to alcohol. A downstream Deep 1D Temporal Residual Convolutional Neural Network architecture will be able to learn specific embedded alcohol-dynamic features in the collected sensing data coming from the GHT25S air-quality sensor of STMicroelectronics. A parallel deep attention-augmented architecture identifies and tracks the salient pedestrians in the driving scenario. A risk assessment system evaluates the sobriety of the driver in case of the presence of salient pedestrians in the driving scene. The collected preliminary results confirmed the effectiveness of the proposed approach.

A Study on Bearing Dynamic Features under the Condition of Multiball–Cage Collision

Cage stability directly affects the dynamic performance of rolling bearing, which, in turn, affects the operating state of rotating equipment. The random collision between the rolling elements and the cage pocket is the main reason for cage instability. In this paper, from the perspective of the relative sliding velocity between the rolling elements and the bearing raceway, the interactions of the rolling elements and the cage pockets were analyzed, and the four zones with different collision features were defined. On this basis, and on the basis of the bearing dynamics model, the interaction of two adjacent rolling elements and the cage pockets in the a’–b’ area is discussed, and the peak impact force of the adjacent two balls and the cage pockets was investigated in terms of the rotation speed, radial load, acceleration/deceleration, and materials. When the ball runs close to the loaded zone, the probability of multiball random collision increases, which leads to an increase in the cage instability. At the entrance of the loaded zone, the peak impact force has the greatest impact on the cage stability during the acceleration process. Compared to the radial load applied to the bearing, the peak impact force is more sensitive to the bearing speed changes. The multiball collision analysis method provides a new idea for the research of cage stability.

Optimal Ambulatory Vital Sign Targets in Pulmonary Hypertension Pregnant Women via Time-Dependent Survival Analysis

Background: Pulmonary hypertension (PH) can cause complications in pregnant women due to significant hemodynamic fluctuation or right heart failure as well as death during pregnancy and postpartum. Those in critical condition would be sent to the intensive care unit (ICU) for observation and treatment. However, evidence to suggest the safe target vital signs is limited and none specific to pregnancy with PH.Methods: This retrospective study of consecutive obstetric patients with PH admitted to ICU of the First Affiliated Hospital of Air Force Military Medical University of China, from January 2011 to May 2020, consisted of 92 cases analyzed using time-dependent Cox regression to consider the dynamic features of vital signs. Results: 7/92 maternal deaths occurred. Most of these deaths occurred within the first three days of admission to the ICU. The vital signs for survival were stable and normal compared to death. Three vital signs were identified as risk factors in the maternal in-hospital mortality model via backward selection: SpO2（HR，0.93；95%CI，0.88-0.97；P=0.003）, heart rate（HR，0.94；95%CI，0.90-0.99；P=0.027）, and mean arterial pressure (MAP) （HR，1.09；95%CI，1.00-1.18；P=0.045）. Log of relative hazard ratios of mortality is linearly negatively related to SpO2 value with a U-shaped correlation with heart rate and MAP (both lower and higher values were associated with high mortality). The optimal range of SpO2 <73%, MAP was 65–95 mmHg, and heart rate was 59–125 beats per minute (bpm). Further exploration showed that the cumulative and the longest consecutive time of abnormal vital signs also affect the outcome. For example, SpO2<73% accumulated for 5 h or continuously up to 2 h increases mortality.Conclusions: Pregnant women with PH who died in the hospital experienced long-term abnormal fluctuations in MAP, heart rate, and SpO2 during ICU stay. Maintaining SpO2>73%, MAP at 65–95mmHg, and heart rate at 59–125 bpm can significantly reduce in-hospital maternal mortality. The effects of the abnormal SpO2, heart rate, and MAP on in-hospital maternal mortality should be combined with the cumulative time and the longest duration.Trial Registry: ChiCTR2100046637.

Mapping the allosteric effects that define functional activity of SARS-CoV-2 specific antibodies

Previous studies on the structural relationship between human antibodies and SARS-CoV-2 have focused on generating static snapshots of antibody complexes with the Spike trimer. However, antibody-antigen interactions are dynamic, with significant binding-induced allosteric effects on conformations of antibody and its target antigen. In this study, we employ hydrogen-deuterium exchange mass spectrometry, in vitro assays, and molecular dynamics simulations to investigate the allosteric perturbations linked to binding events between a group of human antibodies with differential functional activities, and the Spike trimer from SARS-CoV-2. Our investigations have revealed key dynamic features that define weakly or moderately neutralizing antibodies versus those with strong neutralizing activity. These results provide mechanistic insights into the functional modes of human antibodies against COVID-19, and provide a rationale for effective antiviral strategies.TeaserDifferent neutralizing antibodies induce site-specific allosteric effects across SARS-CoV-2 Spike protein

Indonesian Genomic Landscape of Pathogenic Mutation of APC, KRAS, TP53, PIK3CA, and MLH1 in Colorectal Cancer

Background : Knowing colorectal cancer’s heterogeneity and dynamic features, recognizing its biological behaviour requires detailed identification of mutated genes involved. Colorectal cancer (CRC) requires several mutated genes to occur and those are dissimilar in each person hence essential to be discovered in specific population. Until recently, there is no known study describing genomic landscape of CRC in Indonesian population. This study aims to describe profile of pathogenic mutation of APC, TP53, PIK3CA, KRAS, and MLH1 in CRC patients treated at 3 different hospitals in Jakarta. Methods : This is a descriptive study conducted on CRC patients who underwent neoadjuvant, surgical, and adjuvant therapy at RSCM, RSKJ, and MRCCC in 2017-2018. DNA analysis was performed using next-generation sequencing and aligned against GRCh38. Pathogenic variant was identified using ACMG classification and FATHMM score. Data related to behaviour and survival were collected from medical records. Results : There were total 22 subjects in which APC, TP53, and PIKCA were mutated. KRAS mutation occurred in 64%, while MLH1 in 45%. Five types of mutation were identified, including nonsense, missense, frameshift, splice-site, and silent mutation. There are 4 groups of co-occurring mutations, which are APC, TP53, PIK3CA (triple mutation/TM) alone; TM+KRAS; TM+MLH1; and TM+KRAS+MLH1, presenting different nature and survival. Conclusion : Indonesia having various ethnicities with diverse diet and lifestyle has distinct profile of pathogenic mutation presenting mostly with locally-advanced stage with various outcome and survival rate.

Internet Rumors During the COVID-19 Pandemic: Dynamics of Topics and Public Psychologies

The capturing of social opinions, especially rumors, is a crucial issue in digital public health. With the outbreak of the COVID-19 pandemic, the discussions of related topics have increased exponentially in social media, with a large number of rumors on the Internet, which highly impede the harmony and sustainable development of society. As human health has never suffered a threat of this magnitude since the Internet era, past studies have lacked in-depth analysis of rumors regarding such a globally sweeping pandemic. This text-based analysis explores the dynamic features of Internet rumors during the COVID-19 pandemic considering the progress of the pandemic as time-series. Specifically, a Latent Dirichlet Allocation (LDA) model is used to extract rumor topics that spread widely during the pandemic, and the extracted six rumor topics, i.e., “Human Immunity,” “Technology R&amp;D,” “Virus Protection,” “People's Livelihood,” “Virus Spreading,” and “Psychosomatic Health” are found to show a certain degree of concentrated distribution at different stages of the pandemic. Linguistic Inquiry and Word Count (LIWC) is used to statistically test the psychosocial dynamics reflected in the rumor texts, and the results show differences in psychosocial characteristics of rumors at different stages of the pandemic progression. There are also differences in the indicators of psychosocial characteristics between truth and disinformation. Our results reveal which topics of rumors and which psychosocial characteristics are more likely to spread at each stage of progress of the pandemic. The findings contribute to a comprehensive understanding of the changing public opinions and psychological dynamics during the pandemic, and also provide reference for public opinion responses to major public health emergencies that may arise in the future.

A statistical technique for determination of withdrawal of northeast monsoon over coastal Tamilnadu

The withdrawal dates of northeast monsoon over coastal Tamil Nadu for the 90-year period (1901-90) have been objectively derived. The methodology of determination was generally based on an index based on the spatial distribution of daily rainfall over stations of coastal Tamil Nadu, over a 5-day pentad for the six month period, September- February. The normal withdrawal date thus obtained was 27 December with a standard deviation of 13.6 days and range 23 November-28 January. The duration of northeast monsoon was distributed with mean 67.5 days, standard deviation 14.9 days and range 26-102 days. During 36.7 % of years the withdrawal spilled over to January of next year. The daily normal rainfall and its difference filter have been discussed with reference to the normal date of withdrawal. The average decrease of rainfall at the time of withdrawal has been derived by application of superposed epoch analysis. It has further been shown that during years when the withdrawal took place in January the intensity of northeast monsoon prior to withdrawal was as intense as in years when withdrawal occurred in December. A few cases of northeast monsoon withdrawal have been illustrated with diagrams. As no definite dynamic or thermodynamic features could be uniquely identified which are associated with the withdrawal, this technique is basically statistical, considering the behaviour of the daily normal rainfall as the sole criterion. Unique thermodynamic and dynamic features are not identifiable which are associated with the withdrawal of northeast monsoon over coastal Tamilnadu.

Novel Machine Learning for Human Actions Classification Using Histogram of Oriented Gradients and Sparse Representation

Recognition of human actions is a trending research topic as it can be used for crucial medical applications like life care and healthcare. In this research, we propose a novel machine learning algorithm for the classification of human actions based on sparse representation theory. In the proposed framework, the input videos are initially partitioned into several temporal segments of a predefined length. From these temporal segments, the key-cuboids are then obtained. These cuboids are obtained based on the locations having maximum variation in orientation. From these regions, key-cuboids are extracted. From the key-cuboids, Histogram of Oriented Gradient (HOG) features are extracted. This new descriptor has the capability to express the dynamic features in the action videos. Using these features, a single shared dictionary is created from the videos belonging to different classes using K-Singular Value Decomposition (K-SVD) algorithm. This dictionary has the combined features of all the action classes. This shared dictionary is generated during the training phase. During the testing phase, the features belonging to a test class is classified using a novel Sparse Representation Modeling based Action Recognition (SRMAR) Algorithm using Orthogonal Matching Pursuit (OMP) and the shared dictionary. The proposed framework was evaluated using popular benchmark action recognition datasets like KTH dataset, Olympic dataset and the Hollywood dataset. The results obtained using these datasets were represented in the form of a confusion matrix. Evaluation was performed using metrics like overall classification accuracy, specificity, precision, recall and F-score that were obtained from the confusion matrix. This system achieved a high specificity of about 99.52%, 99.16% and 96.15% for the KTH dataset, Olympic dataset and the Hollywood datasets, respectively. Similarly, the proposed framework attained very good precision of 97.64%, 90.46% and 73.39% for the KTH dataset, Olympic dataset and the Hollywood datasets, respectively. Also, the average value of recall achieved was 97.58%, 90.86% and 74.09% for the KTH dataset, Olympic dataset and the Hollywood datasets, respectively. It was also observed that the proposed machine learning algorithm achieved outstanding results compared to the existing state-of-the-art human action recognition frameworks in the literature.