Manual Judgments Research Articles

Research on Liability Identification System of Road Traffic Accident

<p>With the rapid development of social economy and the increasing improvement of people’s living standards, car ownership is increasing exponentially, and road traffic accidents occur frequently. The identification of liability of traffic accident is an important problem in accident handling, which relates to the life and property interests of the litigants concerned. At present, the identification of liability of traffic accident basically depends on the manual judgment of law enforcement department, which is influenced by the experience and human feelings of law enforcement personnel, and has certain randomness and uncertainty. This paper proposes a traffic accident intelligent responsibility identification system based on Case-based Reasoning and D-S Evidence Theory. In the process of liability identification, Case-based Reasoning and D-S Evidence Theory are combined to analyze cases in case database. According to the case similarity, the basic probability of the litigants bearing the main liability is fused to obtain the probability interval of each litigant bearing the main liability, so as to determine the main liability person of the accident. By testing in real cases, the method proposed in this paper has achieved outstanding effect in identifying the main liability of traffic accidents.</p> <p> </p>

Dynamic feeding method for aquaculture fish using multi-task neural network

In recirculating aquaculture system (RAS), fish feeding is the most important part in production management, which is not only related to economic benefits, but also the key to ensure fish welfare and increase production. At present, in RAS, fish are basically fed either artificially or automatically (quantitatively supply feed at definite time), which can easily result in under-feeding or over-feeding of fish. Therefore, there is an urgent to develop an intelligent method that realizes appropriate feeding according to the actual demands of fish. This research attempts to explore a fish dynamic feeding method based on the multi-task network to meet the automatic adjustment of both the feeding intervals (the time intervals between feeding points in repeated feeding in a single-round) and feeding rates. The specific objectives of this study include two parts: 1) to construct a multi-task network to analyze the feeding activity of cultured fish and monitor the amount of uneaten feed pellets; 2) to design a feeding strategy based on information obtained from the multi-task network that realizes the dynamic adjustment of feeding intervals and the decision of feeding endpoint. The waste of feed pellets can be reduced by dynamically adjusting the feeding intervals, and the under-feeding and over-feeding of fish can be prevented by determining feeding endpoint. The results indicated that the accuracy of feeding activity classification by multi-task network reached 95.44%, and the mean absolute error (MAE) and mean square error (MSE) in uneaten feed pellet counting were 4.80 and 6.75, which indicate that the multi-task network can accurately monitor the fish feeding activity and the amount of uneaten feed pellets. Based on the two monitored information, combined with the feeding strategy, we dynamically adjusted the feeding intervals and determined the feeding endpoint, and then compared the feeding endpoints with manual judgment to verify the feasibility and accuracy of the dynamic feeding method based on the multi-task network. In summary, this research provides a more accurate and efficient solution for the intelligent and precise feeding of cultured fish, and provides the theoretical foundation for the development of intelligent feeding devices.

Accurate Detection of Chylous Blood Levels by Deep Learning

Chylous blood, a kind of abnormal blood with an increasing proportion, should be excluded from unpaid blood donation. Therefore, it is important to determine the level of chylous blood, to judge that whether blood is applicable. In this manuscript, the image was preprocessed by the image acquisition device of the background board, and an improved ResNet-50 neural network model of chylous blood level based on images was established. The first layer of convolution was changed from the 7×7-layer convolution kernel of the traditional ResNet-50 model to 5×5-layer convolution kernel of the improved ResNet-50 model, so as to achieve a more accurate image classification effect of chylous blood level. The number of convolution kernel parameters was 25 channels, which decreased the parameters by about 44.5%, and thus greatly reducing the amount of calculation and improving the ability of the network to distinguish image details. It was found that the average values of precision, sensitivity, and F1-Score of the model were all above 0.95. Furthermore, the accuracy of the 5-level determination of the plasma chylous blood level was above 0.95. Compared with manual judgment, this method significantly improved the efficiency and accuracy of detection.

Magnetic Target Detection Using PointRend-Based Region-Convolutional Neural Network

The improving performance of high-resolution magnetic instruments has propelled the refined detection of small targets of different shapes. The traditional method is to extract target-related features through some empirical formulas, and then classify the targets visually. Such manually designed formulas have limited feature expression ability. Also, the results interpreted by different people may vary. Although machine learning was later applied to the task, which avoided the link of manual judgment, the accuracy and robustness were not strong. In this paper, an end-to-end Region Convolutional Neural Network (R-CNN) is proposed to identify targets with little human intervention. Considering the differences between magnetic signals and natural images in terms of scenario, target and imaging, improvements to R-CNN meta-architecture are required. Specifically, magnetic tensor gradient (MTG) data with grid cells are transformed into 2-D matrix, which is then enhanced by pseudo-color coding for mapping. Given the self-built dataset, we design a Two-stage Fine-grained (TSFG) R-CNN to excavate effective deep-level features of targets. PointRend is used here to predict the high-quality edge segmentation of targets. Experiment results show that the proposed method provides a useful way for the detection of multi-scale, multi-shape and multi-depth magnetic targets, even under the case of magnetic field superimposition.

Traffic incident detection based on a global trajectory spatiotemporal map

Traffic incidents endanger the smooth running of vehicles. Congestion caused by traffic incidents has caused a waste of time and fuel and seriously affected transportation efficiency. At present, most methods use manual judgment or image features to detect traffic incidents, but these methods lack timeliness, leading to secondary incidents. For dangerous road sections such as ramp-free and long downhills, this paper proposes an algorithm to quickly detect traffic incidents based on a spatiotemporal map of vehicle trajectories. First, a vehicle dataset from the monitoring perspective is constructed, and an improved YOLOv4 detection algorithm is used to detect images organized as batches. Based on the detection result, the multi-object tracking method of vehicle speed prediction in key frames is used to obtain the vehicle trajectory. Then according to the vehicle trajectory obtained in a single scene, the vehicle trajectory is reidentified and associated in the continuous monitoring scene to construct a long-distance vehicle trajectory spatiotemporal map. Finally, according to the distribution and generation status of the trajectory in the spatiotemporal map, traffic incidents such as vehicle parking, vehicle speeding, and vehicle congestion are analyzed. Experimental results show that the proposed method greatly increases the speed of vehicle detection and tracking and obtains high mAP, MOTA, and MOTP indicators. The global spatiotemporal map constructed by trajectory reidentification can achieve high detection rates for traffic incidents, reduce the average elapsed time, and avoid the problems of the inaccuracy of analyzing image features.

A Review of Breast Cancer Histopathological Image Classification

Breast Cancer (BC) is the most common malignant tumor for women in the world. Histopathological examination serves as basis for breast cancer diagnosis. Due to the low accuracy of histopathological images through manual judgment, the classification of histopathological images of breast cancer has become a research hotspot in the field of medical image processing. Accurate classification of images can help doctors to properly diagnoses and improve the survival rate of patients. This paper reviews the existing works on histopathological image classification of breast cancer and analysis the advantages and disadvantages of related algorithms. Findings of the histopathological image classification of the Breast Cancer study are drawn, and the possible future directions are also discussed.

Putting humans back in the loop of machine learning in Canadian smart cities

AbstractEven as researchers recognize smart cities as sociotechnological assemblages, the attractiveness of artificial intelligence and machine learning (AI/ML) continues to drive cities towards automating urban process and analysis. Rather than arguing whether researchers should automate their analysis, we are interested in identifying where non‐automated, manual judgement calls are made and the analysis is more subjective than presented. We examine topic modelling, an ML method, in an analysis of applications to a pan‐Canadian smart cities grant competition. We document 11 steps in topic modelling from data collection to interpretation. At each step, including the choice of the topic modelling method, some degree of human intervention is required. We draw on human‐centred ML research to argue for a greater recognition of the role of humans‐as‐researchers to preclude further uncritical adoption of AI/ML to research smart cities.

Comprehensive evaluation method of the poultry house indoor environment based on gray relation analysis and analytic hierarchy process

Due to the antagonism and synergy among environmental factors in the poultry house, the influence process becomes extremely complex. As a result, it is difficult to predict and evaluate the degree of such influence accurately. In this paper, we study the poultry house environment factor and its relationship with poultry production performance, using the gray relation analysis (GRA) to filtrate the main factors that influence the evaluation of the poultry house environment. Put forward using the gray relation degree (GRD) to improve the method for structuring the judgment matrix, and weights are more objective and reasonable. The evaluation index system and evaluation model are constructed through the analytic hierarchy process (AHP). It is expected that the comprehensive evaluation of the indoor environment status of the poultry house can guide the optimization of the environmental control in the poultry house and obtain better production indicators of the poultry. In this study, the experimental broiler house was enclosed in autumn. Because of the ventilation system, the indoor environment is still affected by the outdoor environment. The top 3 in the calculation of weights were outdoor environment (0.4315), indoor temperature (0.2384), and indoor air quality (0.1687), which were consistent with experience. From October 24 to 27, the environmental evaluation values of the experimental broiler house were {2.4367, 2.8149, 2.3857, 2.5669}, that is, the evaluation results were {good, good, good, good}; consistent with the expert manual judgment. The correctness and practicability of the proposed method were verified. This paper provides a scientific basis for environmental evaluation and environmental control in the poultry house.

COVID-19 deep classification network based on convolution and deconvolution local enhancement

Computer Tomography (CT) detection can effectively overcome the problems of traditional detection of Corona Virus Disease 2019 (COVID-19), such as lagging detection results and wrong diagnosis results, which lead to the increase of disease infection rate and prevalence rate. The novel coronavirus pneumonia is a significant difference between the positive and negative patients with asymptomatic infections. To effectively improve the accuracy of doctors' manual judgment of positive and negative COVID-19, this paper proposes a deep classification network model of the novel coronavirus pneumonia based on convolution and deconvolution local enhancement. Through convolution and deconvolution operation, the contrast between the local lesion region and the abdominal cavity of COVID-19 is enhanced. Besides, the middle-level features that can effectively distinguish the image types are obtained. By transforming the novel coronavirus detection problem into the region of interest (ROI) feature classification problem, it can effectively determine whether the feature vector in each feature channel contains the image features of COVID-19. This paper uses an open-source COVID-CT dataset provided by Petuum researchers from the University of California, San Diego, which is collected from 143 novel coronavirus pneumonia patients and the corresponding features are preserved. The complete dataset (including original image and enhanced image) contains 1460 images. Among them, 1022 (70%) and 438 (30%) are used to train and test the performance of the proposed model, respectively. The proposed model verifies the classification precision in different convolution layers and learning rates. Besides, it is compared with most state-of-the-art models. It is found that the proposed algorithm has good classification performance. The corresponding sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and precision are 0.98, 0.96, 0.98, and 0.97, respectively.

Spatio-Temporal Perturbations for Video Attribution

The attribution method provides a direction for interpreting opaque neural networks in a visual way by identifying and visualizing the input regions/pixels that dominate the output of a network. Regarding the attribution method for visually explaining video understanding networks, it is challenging because of the unique spatiotemporal dependencies existing in video inputs and the special 3D convolutional or recurrent structures of video understanding networks. However, most existing attribution methods focus on explaining networks taking a single image as input and a few works specifically devised for video attribution come short of dealing with diversified structures of video understanding networks. In this paper, we investigate a generic perturbation-based attribution method that is compatible with diversified video understanding networks. Besides, we propose a novel regularization term to enhance the method by constraining the smoothness of its attribution results in both spatial and temporal dimensions. In order to assess the effectiveness of different video attribution methods without relying on manual judgement, we introduce reliable objective metrics which are checked by a newly proposed reliability measurement. We verified the effectiveness of our method by both subjective and objective evaluation and comparison with multiple significant attribution methods.

RETRACTED ARTICLE: Teaching management of music academies based on multi-core processors and machine learning

Music colleges are a branch of our country’s colleges and universities, and campus informatization is a new development trend under the rapid development of network information technology today. Informatization is conducive to better campus management and administrative efficiency greatly improves. Compared with ordinary colleges and universities, the development of informatization construction in music and art colleges is relatively backward. Therefore, it is very necessary to customize a complete information campus construction plan specifically for music schools. This article will study the related issues of how to carry out information management in music academies based on the problems and practice of teaching management in music schools. Focus on the embedded multi-core processor storage system and propose new storage architecture based on the original. And through machine learning methods, artificial intelligence software that can recognize music styles is trained. Traditional recognition is relatively dependent on the listener’s experience. Cannot reach a complete standard, the accuracy rate is also unstable. So we use embedded application software under the premise of low power consumption. To achieve a stable and efficient way of listening to replace the traditional manual judgment.

Extra-Thyroid Extension Prediction by Ultrasound Quantitative Method Based on Thyroid Capsule Response Evaluation.

BackgroundThe aim of this study was to assess the interaction between thyroid malignancies and thyroid anterior capsule by ultrasound quantification to determine extra-capsular invasion.Material/MethodsA total of 145 patients preoperatively diagnosed with malignant nodules under the thyroid anterior capsule were selected and routinely examined by ultrasound. The length of the nodules (from the junction of the nodule capsule to the deepest point of the nodule, vertical diameter, V) and the distance between the nodule protruding from thyroid capsule and the highest protruding (ledge length, L) nodule were used to obtain the L/V ratio. These parameters where then used to compare the efficacy of predicting extra-thyroid extension (ETE) between L/V, the aspect ratio of the tumor, and manual judgment.ResultsOut of 145 nodules, there were 63 ETEs and 82 non-ETEs determined by ultrasound. Extra-capsular invasion was associated with L//V ratio, but there was no significant correlation between capsular invasion and AR (aspect ratio), age, location, or presence of clustered calcification. The ability of the ratio of L/V to predict extra-capsular invasion was superior to the predictive ability of the AR ratio. With a Youden index of 0.593, the L/V ratio was 0.2325. The use of the L/V ratio to determine the presence of ETE was superior to subjective visual judgment.ConclusionsThe calculation of L/V ratio by ultrasound could more precisely predict the ETE compared with manual judgment, which indirectly reflects the interaction between thyroid capsule and malignant nodules. The above conclusions need to be confirmed by a range of cases.

Intelligent Detection of Urban Road Underground Targets by Using Ground Penetrating Radar based on Deep Learning

Ground penetrating radar (GPR) is widely used in the field of intelligent road detection because of its non-destructive detection method, which is based on an electromagnetic wave reflection mechanism. However, this method requires large-scale data processing and also relies on manual judgment, which is time-consuming and laborious. To solve this problem, this paper analyzes and studies the GPR images of underground pipelines and uneven settlement of urban roads by means of road surface measurement and laboratory tests. The radar image data set is constructed by collecting radar images and denoising and marking them. Then, Deep Feature Selection Net is adopted to improve the fast region-based convolutional neural network (Faster R-CNN) to enhance the network’s ability to extract features from radar images. Finally, by comparing with the improved faster R-CNN model, it is found that the automatic identification rate of underground pipelines and uneven settlement in urban roads increases, reaching more than 80%.

Bibliometrics and visualization analysis of China's Service Quality Development Trend

From the aspect of bibliometrics and social network analysis, this paper studies the content characteristics and network evolution of service quality in China, form visual analysis charts and summarizes the research hotspots and development trends. Taking CNKI database as the data source, data mining and association analysis of sample data were carried out by means of manual judgment, keyword co-occurrence, bibliometric software BICOMB and Pajek. 872 related literatures were retrieved from 1999 to 2019, and the tendency of the number of published papers in each year was close to the inverted U type. This study expands the application fields of bibliometrics and social network software, and analyzes the development trend of service quality. The theoretical and practical research on service quality has achieved rapid development in the past 20 years. The library service quality and service quality satisfaction have always gotten a high level of attention, and the characteristics of network platforms for service quality have gradually become apparent because of the development of the Internet and computer technology.

Design of Tire Damage Image Recognition System Based on Deep Learning

In view of the problems of low efficiency, prone to misjudgment, and complicated identification levels in the current market, the use of manual judgment methods for the cause of tire damage. A method of tire damage image recognition based on convolutional neural network algorithm is proposed, and an intelligent tire damage image recognition system is designed. Through image processing, feature extraction, classification and identification of the input tire damage image, the computer can automatically output the cause of the damage. The experimental results show the feasibility of this method and also improve the shortcomings of traditional manual judgment of the cause of tire damage.

Non-Line-of-Sight Location With Gauss Filtering Algorithm Based on a Model of Photon Flight

Recently, non-line-of-sight (NLOS) detection based on time of flight (TOF) has been investigated. In order to simulate the NLOS location of a hidden object, we derive the signal scattered by the object and build a model of photon flight based on photon scattering and propagation. To improve the authenticity of the model, the bidirectional reflectance distribution function (BRDF) is used to characterize the scattering process. The Gauss filter is proposed to extract the TOF of the scattering sources of interest without a priori information or manual judgment of the useful scattered signal by filtering the disturbance out of the histogram. The hidden object can then be located by TOF processing. Compared with previous work using a fitting algorithm, the Gauss filtering approach preserves more waveform information and presents improved positioning accuracy and robustness under the influence of noise, which is demonstrated in both simulation and experiment. It is possible to locate a NLOS object automatically through filtering identification of the object signal. The simplicity, high efficiency, and automation of this algorithm make it applicable for tracking a hidden moving object.

An Adaptive Online Blade Health Monitoring Method: From Raw Data to Parameters Identification

Blade tip timing (BTT) methods have been increasingly implemented for blade health monitoring (BHM). However, there are two drawbacks of current signal analysis methods preventing them from applying to online monitoring: first, current online monitoring requires the manual judgment of the resonance region, which is time-consuming. Second, existing BTT resonance signal analysis methods are not suitable for online monitoring. The spectral-analysis-based method presents spectral aliasing, while the computational complexity of the sparse-based method is usually high. In this article, we propose an adaptive online BHM method that includes two steps: automatic resonance region recognition and parameters identification in resonance area. For the former step, we demonstrate different methods for synchronous and asynchronous resonances, use the cross correlation to judge the occurrence of synchronous vibration, and use the linear estimation to determine the appearance of asynchronous vibration. For the latter step, an iterative adaptive least-squares periodogram is adopted for its tradeoff between spectral aliasing and computational complexity. The effectiveness of the above steps is first verified using different simulation data separately. Then, the laboratory data are used to test the effectiveness of the whole method. Finally, the online monitoring function of the proposed method is verified by the engineering data with both synchronous and asynchronous resonances.

A Novel Autoencoder with Dynamic Feature Enhanced Factor for Fault Diagnosis of Wind Turbine

Due to the complicated operating environment and variable operating conditions, wind turbines (WTs) are extremely prone to failure and the frequency of fault increases year by year. Therefore, the solutions of effective condition monitoring and fault diagnosis are urgently demanded. Since the vibration signals contain a lot of health condition information, the fault diagnosis based on vibration signals has received extensive attention and achieved impressive progress. However, in practice, the collected health condition signals are very similar and contain a lot of noise, which makes the fault diagnosis of WTs more challenging. In order to handle this problem, this paper proposes a model called denoising stacked feature enhanced autoencoder with dynamic feature enhanced factor (DSFEAE-DF). Firstly, a feature enhanced autoencoder (FEAE) is constructed through feature enhancement so that the discriminative features can be extracted. Secondly, a feature enhanced factor which is independent of manual judgments is proposed and embedded into the training process. Finally, the DSFEAE-DF, combining one noise adding scheme, stacked FEAEs and dynamic feature enhanced factor, is established. Through experimental comparisons, the superiorities of the proposed DSFEAE-DF are verified.

Classification of EEG signals in epilepsy using a novel integrated TSK fuzzy system

The use of machine learning technology to recognize electrical signals of the brain is becoming increasingly popular. Compared with doctors’ manual judgment, machine learning methods are faster. However, only when its recognition accuracy reaches a high level can it be used in practice. Due to the difference in the data distributions of the training dataset and the test dataset and the lack of training samples, the classification accuracies of general machine learning algorithms are not satisfactory. In fact, among the many machine learning methods used to process epilepsy electroencephalogram (EEG) signals, most are black box methods; however, in medicine, methods with explanatory power are needed. In response to these three challenges, this paper proposes a novel technique based on domain adaptation learning, semi-supervised learning and a fuzzy system. In detail, we use domain adaptation learning to reduce deviation from the data distribution, semi-supervised learning to compensate for the lack of training samples, and the Takagi-Sugen-Kang (TSK) fuzzy system model to improve interpretability. Our experimental results show that the performance of the new method is better than those of most advanced epilepsy classification methods.

Abstract 5100: Optimization of amplicon design for polymerase chain reaction

Abstract Amplicon design is critical important in DNA sequencing technology. Previously the amplicon designer uses exhaustive trial and error method, relying on the individual experience to compare the cost of different amplicon combinations. However, optimization of amplicon design has remained challenging due to the lack of a systematic method. As the magnitude of target DNA region and amplicon combination increases, the complexity of manual calculation and judgment will grow exponentially. In this study, our aim was to design a mathematical model to minimize the cost of amplicon combination. For given amplicon design problem, a linear integer programming model is constructed according to the optimization target and the restrictions. Let’s denote our target regions by R_1⋯ R_M, and amplicon by A_1,⋯, A_N, and conflict value c_ij for the amplicon pair A_i and A_j. The model’s restriction is to find a set of amplicons that could cover the target regions as much as possible, and for each R_i we have at most two A_j which will cover it simultaneously. The model’s optimization target is to minimize the total conflict value. The model is implemented by Python programming and uses Gurobi optimization software. In a project to design amplicon for DNA sequencing, the target regions consist of about 90,000 bits, and there are about 35,000 amplicon available. The above optimization method running on the personal computer with 8GB and 2.2GHz Intel i5-5200 CPU, take about 24 hours and produce total conflict value of 1374. The best trial and error practice took more than one week and produced total conflict value of 4527. Optimization of amplicon design for polymerase chain reaction proves to be more precise and efficient than the exhaustive trial and error method. Using optimization method will help standardize the product design for DNA sequencing. Citation Format: Bolong He, Hao Qin, Fugen Li, Hanyan Yang. Optimization of amplicon design for polymerase chain reaction [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 5100.