11th International Conference of Pattern Recognition Systems (ICPRS 2021)最新文献_第3页

Chest X-ray Classification of Pneumonia and COVID19 Using Modified Capsule Networks 基于改进胶囊网络的肺炎和covid - 19胸片分类

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1438

R. Ghosh

{"title":"Chest X-ray Classification of Pneumonia and COVID19 Using Modified Capsule Networks","authors":"R. Ghosh","doi":"10.1049/icp.2021.1438","DOIUrl":"https://doi.org/10.1049/icp.2021.1438","url":null,"abstract":"Many studies are already done on Deep Learning-based diagnosis, specially using Convolutional Neural Network (CNN), to assist identifying lung disease cases based on radiology imaging. In this study three types of chest X-ray images are taken to be classified by convolutional neural network (CNN), e.g. 1583 normal or healthy chest X-rays, 4273 pneumonia diagnosed chest X-rays and 262 COVID19 diagnosed chest X-ray images. Five various proved architectures (VGG16, VGG19, Xception, InceptionV3, Inception-ResNetV2) are tested on diagnosis of the above classes of X-rays images. Then this above five convolutional architectures are used as feature extractors for a capsule layer of 16 capsule dimension and 4 routings. Total ten CNN architectures are tested to perform the task. The main advantages of capsule networks is that the part-whole relation can be captured through the capsules of consecutive layers. Among the tested main five CNNs VGG16 performs the best with 96.65% accuracy over this task. Among the other five capsulated CNNs VGG16 based capsule network outperforms any other architecture tested with an accuracy of 96.81%. Hopefully the proposed CNN architecture may be an alternative method to diagnose any X-ray classification by providing fast and accurate screening.","PeriodicalId":431144,"journal":{"name":"11th International Conference of Pattern Recognition Systems (ICPRS 2021)","volume":"92 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126825188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Domain Adaption in Sequence-to-Sequence Scene Text Recognition 序列到序列场景文本识别中的领域自适应

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1449

Zheng Li, Joshua Smith, Sujoy Chakraborty

引用次数: 0

Deep models and optimizers for Indian sign language recognition 印度手语识别的深度模型和优化器

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1445

P. Sharma, R. Anand

引用次数: 0

A methodology for procedural piano music composition with mood templates using genetic algorithms 基于遗传算法的情绪模板程序化钢琴音乐创作方法

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1435

L. Rocha de Azevedo Santos, C. N. Silla Jr, Márjory Da Costa-Abreu

引用次数: 2

On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks 非突触反向传播在长期认知网络训练中的性能研究

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1434

Gonzalo N´apoles, Isel Grau, Leonardo Concepci´on, Yamisleydi Salgueiro

{"title":"On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks","authors":"Gonzalo N´apoles, Isel Grau, Leonardo Concepci´on, Yamisleydi Salgueiro","doi":"10.1049/icp.2021.1434","DOIUrl":"https://doi.org/10.1049/icp.2021.1434","url":null,"abstract":"Long-term Cognitive Networks (LTCNs) are recurrent neural networks for modeling and simulation. Such networks can be trained in a synaptic or nonsynaptic mode according to their goal. Nonsynaptic learning refers to adjusting the transfer function parameters while preserving the weights connecting the neurons. In that regard, the Nonsynaptic Backpropagation (NSBP) algorithm has proven successful in training LTCN based models. Despite NSBP's success, a question worthy of investigation is whether the backpropagation process is necessary when training these recurrent neural networks. This paper investigates this issue and presents three nonsynaptic learning methods that modify the original algorithm. In addition, we perform a sensitivity analysis of both the NSBP's hyperparameters and the LTCNs' learnable parameters. The main conclusions of our study are i) the backward process attached to the NSBP algorithm is not necessary to train these recurrent neural systems, and ii) there is a nonsynaptic learnable parameter that does not contribute significantly to the LTCNs' performance.","PeriodicalId":431144,"journal":{"name":"11th International Conference of Pattern Recognition Systems (ICPRS 2021)","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124404515","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Asymmetric U-Net for Brain Tumor Segmentation: Transfer to an independent database 用于脑肿瘤分割的非对称U-Net:转移到一个独立的数据库

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1447

S. R. González, I. Zemmoura, C. Tauber

引用次数: 0

Real Time Architecture for Image De-Hazing 图像去雾的实时架构

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1457

Laiba Khurshid, G. Raja

{"title":"Real Time Architecture for Image De-Hazing","authors":"Laiba Khurshid, G. Raja","doi":"10.1049/icp.2021.1457","DOIUrl":"https://doi.org/10.1049/icp.2021.1457","url":null,"abstract":"Outdoor scenes are mostly affected by bad weather and observed objects in images suffer from poor visibility and contrast. Therefore de-hazing algorithms are utilized to eliminate haze and fog from images. This paper describes the real-time architecture for image de-hazing using dark channel prior method. In proposed architecture, hazy images are read in patch form and minimum intensity pixels are separated from every patch. These pixels of minimal intensity combine to form a dark vector which is used in atmospheric light computation. Atmospheric light value is used in assessment of transmission map. In order to avoid appearing of artifacts in recovered image, fast guided filter is utilized for transmission map refinement and then refined transmission map is used in recovery of scene. The hardware architecture is implemented on Xilinx VIVADO tool using VHDL language. A comparison of simulation results with MATLAB results validates the architecture of image de-hazing. Hardware architecture is synthesized using Xilinx Virtex board, Device XCVU190, Package FLGC2104, and Speed 1800. Results from synthesis show that architecture consumes 7130 slice registers and 48690 LUTs.","PeriodicalId":431144,"journal":{"name":"11th International Conference of Pattern Recognition Systems (ICPRS 2021)","volume":"2021 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129476235","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Color image deblurring and denoising by using bilateral sharpening 利用双边锐化技术对彩色图像进行去模糊和去噪

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1444

P. González, A. Ubilla, F. Tirado, C. Tauber

引用次数: 0

Investigating the use of feature selection techniques for gender prediction systems based on keystroke dynamics 研究基于击键动力学的性别预测系统的特征选择技术的使用

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1446

Tuany M. L. Nascimento, Andrelyne V. M. Oliveira, L. E. A. Santana, M. Da Costa-Abreu

{"title":"Investigating the use of feature selection techniques for gender prediction systems based on keystroke dynamics","authors":"Tuany M. L. Nascimento, Andrelyne V. M. Oliveira, L. E. A. Santana, M. Da Costa-Abreu","doi":"10.1049/icp.2021.1446","DOIUrl":"https://doi.org/10.1049/icp.2021.1446","url":null,"abstract":"Biometric-based solutions keep expanding with new modalities, techniques and systems being proposed every so often. However, the first ones that were used for authentication, such as handwritten signature and keystroke dynamics, continue to be relevant in our digital world, despite their analogical origin. In special, keystroke dynamics has had an increase in popularity with the advent of social networks, making the need to continue to authenticate in desktop or game-based user verification more prevalent and this became an open door to risky situations such as paedophilia, sexual abuse, harassment among others. One of the ways to combat this type of crime is to be able to verify the legitimacy of the gender of the person the user is typing with. Despite the fact that keystroke dynamics is well accepted and reliable, this technique can have far too many attributes to be analysed which can lead to the use of redundant or irrelevant information. Therefore, propose a comparative study between two features selection approaches, hybrid (filter + wrapper) and wrapper. They will be tested by using a genetic algorithm, a particle swarm optimisation, a k-NN, a SVM, and a Naive Bayes as classifiers, as well as, the Correlation and Relief filters. From the results obtained, it can be said that the two proposed hybrid approaches reduce the number of attributes, without negatively impacting the accuracy of the classification, and being less costly than the traditional PSO.","PeriodicalId":431144,"journal":{"name":"11th International Conference of Pattern Recognition Systems (ICPRS 2021)","volume":"2021 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129658840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Fall Detection using Human Skeleton Features 基于人体骨骼特征的跌倒检测

11th International Conference of Pattern Recognition Systems (ICPRS 2021) Pub Date : 1900-01-01 DOI: 10.1049/icp.2021.1465

H. Ramirez, S. Velastín, E. Fàbregas, I. Meza, D. Makris, G. Farías

引用次数: 6