Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33035
Yu Zhang, Zilong Wang, Yongjian Zhu
In order to solve the problem of uneven density and the low classification accuracy of 3D point cloud, a 3D point cloud classification method fuses multi-attention machine is proposed. It is principally based on the traditional point cloud dynamic graph convolution classification network, into multiple attention mechanisms, including self-attention, spatial attention and channel attention mechanisms. The self-attention mechanism can reduce the dependence on irrelevant points while aligning point clouds, and input the processed point cloud into the classification network. Then the missing geometric information in the classification network is compensated by the integration of spatial and channel attention mechanisms. The experimental results on the public data set ModelNet40 indicate that compared with the DGCNN classification network, the improved network model improves the classification accuracy of the data set by 0.5 % and the average accuracy by 0.9 %. Meantime, the classification accuracy outstrips other contrast classification algorithms.
{"title":"3D Point Cloud Classification Method Based on Multiple Attention Mechanism and Dynamic Graph Convolution","authors":"Yu Zhang, Zilong Wang, Yongjian Zhu","doi":"10.5755/j01.itc.52.3.33035","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33035","url":null,"abstract":"In order to solve the problem of uneven density and the low classification accuracy of 3D point cloud, a 3D point cloud classification method fuses multi-attention machine is proposed. It is principally based on the traditional point cloud dynamic graph convolution classification network, into multiple attention mechanisms, including self-attention, spatial attention and channel attention mechanisms. The self-attention mechanism can reduce the dependence on irrelevant points while aligning point clouds, and input the processed point cloud into the classification network. Then the missing geometric information in the classification network is compensated by the integration of spatial and channel attention mechanisms. The experimental results on the public data set ModelNet40 indicate that compared with the DGCNN classification network, the improved network model improves the classification accuracy of the data set by 0.5 % and the average accuracy by 0.9 %. Meantime, the classification accuracy outstrips other contrast classification algorithms.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134886918","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33042
L. Antony Rosewelt, D. Naveen Raju, E. Sujatha
Customer reviews are playing an important role in e-commerce for increasing sales by knowing the customer’s purchase pattern and expectations. The reviews that are collected after completing their purchase reflect the quality and services in e-commerce. The user’s reviews are characterized and categorized through sentiment and semantic analysis. Moreover, the sentiment and semantic classification processes are also performed to predict the user’s purchase patterns and liked products. However, the available classification is not able to predict the user’s purchase patterns. In this paper, we propose a new Product Recommendation System (PRS) to predict the appropriate product for users based on their purchase behavior and pattern. The proposed recommendation system incorporates the standard data preprocessing tasks like tokenization process, Parts of Speech (PoS) tagging process, and parsing, a new sentiment and semantic score calculation procedure, and a new feature optimization technique called the Weighted Aquila Optimization Method (WAOM). Moreover, the sentiment and semantic classification processes are performed by applying a General Regression Neural Network with the incorporation of fuzzy temporal features (FTGRNN) and obtaining better classification results. The newly developed PRS is evaluated by conducting experiments in this work and also proved as superior than other systems available in this direction in terms of prediction accuracy, precision, recall, serendipity and nDCG.
{"title":"A New Sentiment and Fuzzy Aware Product Recommendation System Using Weighted Aquila Optimization and GRNN in e-Commerce","authors":"L. Antony Rosewelt, D. Naveen Raju, E. Sujatha","doi":"10.5755/j01.itc.52.3.33042","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33042","url":null,"abstract":"Customer reviews are playing an important role in e-commerce for increasing sales by knowing the customer’s purchase pattern and expectations. The reviews that are collected after completing their purchase reflect the quality and services in e-commerce. The user’s reviews are characterized and categorized through sentiment and semantic analysis. Moreover, the sentiment and semantic classification processes are also performed to predict the user’s purchase patterns and liked products. However, the available classification is not able to predict the user’s purchase patterns. In this paper, we propose a new Product Recommendation System (PRS) to predict the appropriate product for users based on their purchase behavior and pattern. The proposed recommendation system incorporates the standard data preprocessing tasks like tokenization process, Parts of Speech (PoS) tagging process, and parsing, a new sentiment and semantic score calculation procedure, and a new feature optimization technique called the Weighted Aquila Optimization Method (WAOM). Moreover, the sentiment and semantic classification processes are performed by applying a General Regression Neural Network with the incorporation of fuzzy temporal features (FTGRNN) and obtaining better classification results. The newly developed PRS is evaluated by conducting experiments in this work and also proved as superior than other systems available in this direction in terms of prediction accuracy, precision, recall, serendipity and nDCG.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134887049","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33583
Haolin Yin, Baoquan Li, Hai Zhu, Lintao Shi
In this paper, an autonomous navigation strategy is proposed for unmanned aerial vehicles (UAVs) based on consideration of dynamic sampling and field of view (FOV). Compare to search-based motion planning, sampling-based kinodynamic planning schemes can often find feasible trajectories in complex environments. Specifically, a global trajectory is first generated with physical information, and an expansion algorithm is constructed regarding to kinodynamic rapidly-exploring random tree* (KRRT*). Then, a KRRT* expansion strategy is designed to find local collision-free trajectories. In trajectory optimization, bending radius, collision risk function, and yaw angle penalty term are defined by taking into account onboard sensor FOV and potentialrisk. Then, smooth and dynamic feasible terms are penalized based on initial trajectory generation. Trajectories are refined by time reallocation, and weights are solved by optimization. Effectiveness of the proposed strategy is demonstrated by both simulation and experiment.
{"title":"Kinodynamic RRT* Based UAV Optimal State Motion Planning with Collision Risk Awareness","authors":"Haolin Yin, Baoquan Li, Hai Zhu, Lintao Shi","doi":"10.5755/j01.itc.52.3.33583","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33583","url":null,"abstract":"In this paper, an autonomous navigation strategy is proposed for unmanned aerial vehicles (UAVs) based on consideration of dynamic sampling and field of view (FOV). Compare to search-based motion planning, sampling-based kinodynamic planning schemes can often find feasible trajectories in complex environments. Specifically, a global trajectory is first generated with physical information, and an expansion algorithm is constructed regarding to kinodynamic rapidly-exploring random tree* (KRRT*). Then, a KRRT* expansion strategy is designed to find local collision-free trajectories. In trajectory optimization, bending radius, collision risk function, and yaw angle penalty term are defined by taking into account onboard sensor FOV and potentialrisk. Then, smooth and dynamic feasible terms are penalized based on initial trajectory generation. Trajectories are refined by time reallocation, and weights are solved by optimization. Effectiveness of the proposed strategy is demonstrated by both simulation and experiment.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134961173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33719
N. Adhithyaa, A. Tamilarasi, D. Sivabalaselvamani, L. Rahunathan
Accidents due to driver drowsiness are observed to be increasing at an alarming rate across all countries and it becomes necessary to identify driver drowsiness to reduce accident rates. Researchers handled many machine learning and deep learning techniques especially many CNN variants created for drowsiness detection, but it is dangerous to use in real time, as the design fails due to high computational complexity, low evaluation accuracies and low reliability. In this article, we introduce a multistage adaptive 3D-CNN model with multi-expressive features for Driver Drowsiness Detection (DDD) with special attention to system complexity and performance. The proposed architecture is divided into five cascaded stages: (1) A three level Convolutional Neural Network (CNN) for driver face positioning (2) 3D-CNN based Spatio-Temporal (ST) Learning to extract 3D features from face positioned stacked samples. (3) State Understanding (SU) to train 3D-CNN based drowsiness models (4) Feature fusion using ST and SU stages (5) Drowsiness Detection stage. The Proposed system extract ST values from the face positioned images and then merges it with SU results from each state understanding sub models to create conditional driver facial features for final Drowsiness Detection (DD) model. Final DD Model is trained offline and implemented in online, results show the developed model performs well when compared to others and additionally capable of handling Indian conditions. This method is applied (Trained and Evaluated) using two different datasets, Kongu Engineering College Driver Drowsiness Detection (KEC-DDD) own dataset and National Tsing Hua University Driver Drowsiness Detection (NTHU-DDD) Benchmark Dataset. The proposed system trained with KEC-DDD dataset produces accuracy of 77.45% and 75.91% using evaluation set of KEC-DDD and NTHU-DDD dataset and capable to detect driver drowsiness from 256×256 resolution images at 39.6 fps at an average of 400 execution seconds.
{"title":"Face Positioned Driver Drowsiness Detection Using Multistage Adaptive 3D Convolutional Neural Network","authors":"N. Adhithyaa, A. Tamilarasi, D. Sivabalaselvamani, L. Rahunathan","doi":"10.5755/j01.itc.52.3.33719","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33719","url":null,"abstract":"Accidents due to driver drowsiness are observed to be increasing at an alarming rate across all countries and it becomes necessary to identify driver drowsiness to reduce accident rates. Researchers handled many machine learning and deep learning techniques especially many CNN variants created for drowsiness detection, but it is dangerous to use in real time, as the design fails due to high computational complexity, low evaluation accuracies and low reliability. In this article, we introduce a multistage adaptive 3D-CNN model with multi-expressive features for Driver Drowsiness Detection (DDD) with special attention to system complexity and performance. The proposed architecture is divided into five cascaded stages: (1) A three level Convolutional Neural Network (CNN) for driver face positioning (2) 3D-CNN based Spatio-Temporal (ST) Learning to extract 3D features from face positioned stacked samples. (3) State Understanding (SU) to train 3D-CNN based drowsiness models (4) Feature fusion using ST and SU stages (5) Drowsiness Detection stage. The Proposed system extract ST values from the face positioned images and then merges it with SU results from each state understanding sub models to create conditional driver facial features for final Drowsiness Detection (DD) model. Final DD Model is trained offline and implemented in online, results show the developed model performs well when compared to others and additionally capable of handling Indian conditions. This method is applied (Trained and Evaluated) using two different datasets, Kongu Engineering College Driver Drowsiness Detection (KEC-DDD) own dataset and National Tsing Hua University Driver Drowsiness Detection (NTHU-DDD) Benchmark Dataset. The proposed system trained with KEC-DDD dataset produces accuracy of 77.45% and 75.91% using evaluation set of KEC-DDD and NTHU-DDD dataset and capable to detect driver drowsiness from 256×256 resolution images at 39.6 fps at an average of 400 execution seconds.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134886923","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.34032
R. Murugesan, K. Devaki
The classic feature extraction techniques used in recent research on computer-aided diagnosis (CAD) of liver cancer have several disadvantages, including duplicated features and substantial computational expenses. Modern deep learning methods solve these issues by implicitly detecting complex structures in massive quantities of healthcare image data. This study suggests a unique bio-inspired deep-learning way for improving liver cancer prediction outcomes. Initially, a novel semantic segmentation technique known as UNet++ is proposed to extract liver lesions from computed tomography (CT) images. Second, a hybrid approach that combines the Chaotic Cuckoo Search algorithm and AlexNet is indicated as a feature extractor and classifier for liver lesions. LiTS, a freely accessible database that contains abdominal CT images, was employed for liver tumor diagnosis and investigation. The segmentation results were evaluated using the Dice similarity coefficient and Correlation coefficient. The classification results were assessed using Accuracy, Precision, Recall, F1 Score, and Specificity. Concerning the performance metrics such as accuracy, precision, and recall, the recommended method performs better than existing algorithms producing the highest values such as 99.2%, 98.6%, and 98.8%, respectively.
{"title":"Liver Lesion Detection Using Semantic Segmentation and Chaotic Cuckoo Search Algorithm","authors":"R. Murugesan, K. Devaki","doi":"10.5755/j01.itc.52.3.34032","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.34032","url":null,"abstract":"The classic feature extraction techniques used in recent research on computer-aided diagnosis (CAD) of liver cancer have several disadvantages, including duplicated features and substantial computational expenses. Modern deep learning methods solve these issues by implicitly detecting complex structures in massive quantities of healthcare image data. This study suggests a unique bio-inspired deep-learning way for improving liver cancer prediction outcomes. Initially, a novel semantic segmentation technique known as UNet++ is proposed to extract liver lesions from computed tomography (CT) images. Second, a hybrid approach that combines the Chaotic Cuckoo Search algorithm and AlexNet is indicated as a feature extractor and classifier for liver lesions. LiTS, a freely accessible database that contains abdominal CT images, was employed for liver tumor diagnosis and investigation. The segmentation results were evaluated using the Dice similarity coefficient and Correlation coefficient. The classification results were assessed using Accuracy, Precision, Recall, F1 Score, and Specificity. Concerning the performance metrics such as accuracy, precision, and recall, the recommended method performs better than existing algorithms producing the highest values such as 99.2%, 98.6%, and 98.8%, respectively.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134886682","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33320
K. Venu, P. Natesan
Brain-Computer Interface (BCI) is a technology in which Electroencephalogram (EEG) signals are utilized to create a link between a person’s mental state and a computer-based signal processing system that decodes the signals without needing muscle movement. The mental process of picturing the movement of a body component without actually moving that body part is known as Motor Imagery (MI). MI BCI is a Motor Imagery-based Brain-Computer Interface that allows patients with motor impairments to interact with their environment by operating robotic prostheses, wheelchairs, and other equipment. Feature extraction and classification are essential parts of the EEG signal processing for MI BCI. In this work, Whales Optimization Algorithm with an Improved Mutualism Phase is proposed to find the optimal Convolutional Neural Network architecture for the classification of motor imagery tasks with high accuracy and less computational complexity. The Neurosky and BCI IV 2a datasets were used to evaluate the proposed methodology. Experiments demonstrate that the suggested technique outperforms other competing methods regarding classification accuracy values at 94.1% and 87.7% for the Neurosky and BCI datasets, respectively.
脑机接口(BCI)是一种利用脑电图(EEG)信号在人的精神状态和基于计算机的信号处理系统之间建立联系的技术，该系统无需肌肉运动即可解码信号。在不实际移动身体部位的情况下，想象身体某个部位运动的心理过程被称为运动想象(MI)。MI BCI是一种基于运动图像的脑机接口，它允许运动障碍患者通过操作机器人假肢、轮椅和其他设备与他们的环境进行互动。特征提取和分类是脑电信号处理的重要组成部分。本文提出了一种改进的互共生阶段的whale优化算法，以寻找最优的卷积神经网络架构，用于高精度和低计算复杂度的运动图像任务分类。使用Neurosky和BCI IV 2a数据集来评估所提出的方法。实验表明，在Neurosky和BCI数据集上，该方法的分类准确率分别为94.1%和87.7%，优于其他竞争方法。
{"title":"Optimized Deep Learning Model Using Modified Whale’s Optimization Algorithm for EEG Signal Classification","authors":"K. Venu, P. Natesan","doi":"10.5755/j01.itc.52.3.33320","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33320","url":null,"abstract":"Brain-Computer Interface (BCI) is a technology in which Electroencephalogram (EEG) signals are utilized to create a link between a person’s mental state and a computer-based signal processing system that decodes the signals without needing muscle movement. The mental process of picturing the movement of a body component without actually moving that body part is known as Motor Imagery (MI). MI BCI is a Motor Imagery-based Brain-Computer Interface that allows patients with motor impairments to interact with their environment by operating robotic prostheses, wheelchairs, and other equipment. Feature extraction and classification are essential parts of the EEG signal processing for MI BCI. In this work, Whales Optimization Algorithm with an Improved Mutualism Phase is proposed to find the optimal Convolutional Neural Network architecture for the classification of motor imagery tasks with high accuracy and less computational complexity. The Neurosky and BCI IV 2a datasets were used to evaluate the proposed methodology. Experiments demonstrate that the suggested technique outperforms other competing methods regarding classification accuracy values at 94.1% and 87.7% for the Neurosky and BCI datasets, respectively.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134961019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33701
Yu Hao, Huimin Du, Meiwen Mao, Ying Liu, Jiulun Fan
Traditional detect and count strategy can’t well handle the extremely crowded footage in computer vision-based counting task. In recent years, deep learning approaches have been widely explored to tackle this challenge. By regressing visual features to density map, the total crowd number can be predicted while avoids the detection of their actual positions. Efforts of improving performance distribute at various phases of the detecting pipeline, such as feature extraction and eliminating deviation of regressed density map etc. In this article, we conduct a thorough review on the most representative and state-of-the-art techniques. The efforts are systematically categorized into three topics: the evolving of front-end network, the handling of unbalanced density map prediction, and the selection of loss function. After the evaluation of most significant techniques, innovations of the state-of-the-art are inspected in detail to analyze specific reasons to achieve high performances. As conclusion, possible directions of enhancement are discussed to provide insights of future research.
{"title":"A Survey on Regression-Based Crowd Counting Techniques","authors":"Yu Hao, Huimin Du, Meiwen Mao, Ying Liu, Jiulun Fan","doi":"10.5755/j01.itc.52.3.33701","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33701","url":null,"abstract":"Traditional detect and count strategy can’t well handle the extremely crowded footage in computer vision-based counting task. In recent years, deep learning approaches have been widely explored to tackle this challenge. By regressing visual features to density map, the total crowd number can be predicted while avoids the detection of their actual positions. Efforts of improving performance distribute at various phases of the detecting pipeline, such as feature extraction and eliminating deviation of regressed density map etc. In this article, we conduct a thorough review on the most representative and state-of-the-art techniques. The efforts are systematically categorized into three topics: the evolving of front-end network, the handling of unbalanced density map prediction, and the selection of loss function. After the evaluation of most significant techniques, innovations of the state-of-the-art are inspected in detail to analyze specific reasons to achieve high performances. As conclusion, possible directions of enhancement are discussed to provide insights of future research.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134961160","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.32258
Leihong Zhang, Zhaoyuan Ji, Runchu Xu, Dawei Zhang
The detection of salient objects in foggy scenes is an important research component in many practical applications such as action recognition, target tracking and pedestrian re-identification. To facilitate saliency detection in foggy scenes, this paper explores two issues. The construction of dataset for foggy weather conditions and implementation scheme for foggy weather saliency detection. Firstly, a foggy sky image synthesis method is designed based on the atmospheric scattering model, and a saliency detection dataset applicable to foggy sky is constructed. Secondly, we compare the current classification networks and adopt resnet50, which has the highest classification accuracy, as the backbone network of the classification module, and classify the foggy sky images into three levels, namely fogless, light fog and dense fog, according to different concentrations. Then, Residual Refinement Network (R2Net) was selected to train and test the classified images. Horizontal and vertical flipping and image cropping were used to enhance the training set to relieve over-fitting. The accuracy of the network model was improved by using Adam as the optimizer. Experimental results show that for the detection of fogless images, our method is almost on par with state-of-the-art, and performs well for both light and dense fog images. Our method has good adaptability, accuracy and robustness.
{"title":"Saliency Detection Algorithm for Foggy Images Based on Deep Learning","authors":"Leihong Zhang, Zhaoyuan Ji, Runchu Xu, Dawei Zhang","doi":"10.5755/j01.itc.52.3.32258","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.32258","url":null,"abstract":"The detection of salient objects in foggy scenes is an important research component in many practical applications such as action recognition, target tracking and pedestrian re-identification. To facilitate saliency detection in foggy scenes, this paper explores two issues. The construction of dataset for foggy weather conditions and implementation scheme for foggy weather saliency detection. Firstly, a foggy sky image synthesis method is designed based on the atmospheric scattering model, and a saliency detection dataset applicable to foggy sky is constructed. Secondly, we compare the current classification networks and adopt resnet50, which has the highest classification accuracy, as the backbone network of the classification module, and classify the foggy sky images into three levels, namely fogless, light fog and dense fog, according to different concentrations. Then, Residual Refinement Network (R2Net) was selected to train and test the classified images. Horizontal and vertical flipping and image cropping were used to enhance the training set to relieve over-fitting. The accuracy of the network model was improved by using Adam as the optimizer. Experimental results show that for the detection of fogless images, our method is almost on par with state-of-the-art, and performs well for both light and dense fog images. Our method has good adaptability, accuracy and robustness.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134960200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33155
Sha Ji, Chengde Lin
A human motion pattern recognition algorithm based on Nano-sensor and deep learning is studied to recognize human motion patterns in real time and with high accuracy. First, human motion data are collected by micro electro mechanical system, and the noise in such data is filtered by smoothing filtering method to obtain high-quality motion data. Second, key time-domain features are extracted from high-quality motion data. Finally, after fusing and processing the key time-domain features, it is input into the deep long and short-term memory (LSTM) neural network to build a deep LSTM human motion pattern recognition model and complete human motion pattern recognition. The results show that the proposed algorithm can realize the recognition of various motion patterns with high accuracy of data acquisition, the average recognition accuracy is 94.8%, the average recall reaches 89.7%, and the F1 score of the algorithm are high, and the recognition time consuming is short, which can realize accurate and efficient human motion pattern recognition and provide guarantee for effective monitoring of the target human motion health.
{"title":"Human Motion Pattern Recognition Based on Nano-sensor and Deep Learning","authors":"Sha Ji, Chengde Lin","doi":"10.5755/j01.itc.52.3.33155","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33155","url":null,"abstract":"A human motion pattern recognition algorithm based on Nano-sensor and deep learning is studied to recognize human motion patterns in real time and with high accuracy. First, human motion data are collected by micro electro mechanical system, and the noise in such data is filtered by smoothing filtering method to obtain high-quality motion data. Second, key time-domain features are extracted from high-quality motion data. Finally, after fusing and processing the key time-domain features, it is input into the deep long and short-term memory (LSTM) neural network to build a deep LSTM human motion pattern recognition model and complete human motion pattern recognition. The results show that the proposed algorithm can realize the recognition of various motion patterns with high accuracy of data acquisition, the average recognition accuracy is 94.8%, the average recall reaches 89.7%, and the F1 score of the algorithm are high, and the recognition time consuming is short, which can realize accurate and efficient human motion pattern recognition and provide guarantee for effective monitoring of the target human motion health.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"2012 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134960369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-26DOI: 10.5755/j01.itc.52.3.33360
Shuo Yin, Youwei Gao, Shuai Nie, Junbao Li
In financial big data field, most existing work of stock prediction has focused on the prediction of a single stock trend. However, it is challenging to predict a stock price series due to its drastic volatility. While the stock sector is a group of stocks belonging to the same sector, and the stock sector index is the weighted sum of the prices of all the stocks in the sector. Therefore the trend of stock sector is more stable and more feasible to predict than that of a single stock. In this paper, we propose a new method named Stock Sector Trend Prediction (SSTP) to solve the problem of predicting stock sector trend. In SSTP method, we adopt the Relative Price Strength (RPS) to describe the trend of the stock sector, which is the relative rank of stock sector trend. In order to learn the intrinsic probability distribution of the stock sector index series, we construct the multi-scale RPS time series and build multiple independent fully-connected stock sector relation graphs based on the real relationship among stock sectors. Then, we propose a Temporal-spatial Network (TSN) to extract the temporal features from the multi-scale RPS series and the spatial features from the stock sector relation graphs. Finally, the TSN predicts and ranks the trends of the stock sector trend with the temporal-spatial features. The experimental results on the real-world dataset validate the effectiveness of the proposed SSTP method for the stock sector trend prediction.
{"title":"SSTP: Stock Sector Trend Prediction with Temporal-Spatial Network","authors":"Shuo Yin, Youwei Gao, Shuai Nie, Junbao Li","doi":"10.5755/j01.itc.52.3.33360","DOIUrl":"https://doi.org/10.5755/j01.itc.52.3.33360","url":null,"abstract":"In financial big data field, most existing work of stock prediction has focused on the prediction of a single stock trend. However, it is challenging to predict a stock price series due to its drastic volatility. While the stock sector is a group of stocks belonging to the same sector, and the stock sector index is the weighted sum of the prices of all the stocks in the sector. Therefore the trend of stock sector is more stable and more feasible to predict than that of a single stock. In this paper, we propose a new method named Stock Sector Trend Prediction (SSTP) to solve the problem of predicting stock sector trend. In SSTP method, we adopt the Relative Price Strength (RPS) to describe the trend of the stock sector, which is the relative rank of stock sector trend. In order to learn the intrinsic probability distribution of the stock sector index series, we construct the multi-scale RPS time series and build multiple independent fully-connected stock sector relation graphs based on the real relationship among stock sectors. Then, we propose a Temporal-spatial Network (TSN) to extract the temporal features from the multi-scale RPS series and the spatial features from the stock sector relation graphs. Finally, the TSN predicts and ranks the trends of the stock sector trend with the temporal-spatial features. The experimental results on the real-world dataset validate the effectiveness of the proposed SSTP method for the stock sector trend prediction.","PeriodicalId":54982,"journal":{"name":"Information Technology and Control","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134886919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：