Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100463
Yehya Bouzeraa, Nardjas Bouchemal, Nada Zendaoui
Due to climate changes, the world has experienced in recent years violent natural disasters such as fires, floods and typhoons. Since Disasters are difficult to cope with and stop, researchers have focused on the pre-disaster management phase to reduce damage. Pre-disaster management is the first phase in a disaster management cycle, it uses available data collected from different resources, to detect and predict disaster cases in order to give rescue teams more time to prepare and make decisions.With development of modern technologies and equipment many systems and approaches were proposed based on this development to improve the effectiveness and performance of disaster management and early warning systems. This paper aims to provide an overview of last years studies, focusing on new technologies (IoT, Machine learning, Big Data) for pre-disaster management.
{"title":"Pre-disaster Management based Machine Learning, IoT and Big Data: Survey and future direction","authors":"Yehya Bouzeraa, Nardjas Bouchemal, Nada Zendaoui","doi":"10.1109/NTIC55069.2022.10100463","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100463","url":null,"abstract":"Due to climate changes, the world has experienced in recent years violent natural disasters such as fires, floods and typhoons. Since Disasters are difficult to cope with and stop, researchers have focused on the pre-disaster management phase to reduce damage. Pre-disaster management is the first phase in a disaster management cycle, it uses available data collected from different resources, to detect and predict disaster cases in order to give rescue teams more time to prepare and make decisions.With development of modern technologies and equipment many systems and approaches were proposed based on this development to improve the effectiveness and performance of disaster management and early warning systems. This paper aims to provide an overview of last years studies, focusing on new technologies (IoT, Machine learning, Big Data) for pre-disaster management.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116983756","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}
In recent years, E-learning technologies have altered the way we teach and learn, making it an intriguing research topic for enhancing education. A key component of these systems is the ability to tailor the learning experience to the needs of the individual student. According to researches, modeling student profiles with an ontology is quite relevant. However, the ontology must consider every aspect of learner representation. Therefore, there is an urgent need for new comprehensive information to improve the learner profile. In this paper, we propose a semantic approach to define an ontology of learner profiles. In addition, a learning style prediction system based on machine learning techniques is developed. Empirical results show a promising gain in performance for learning style prediction systems.
{"title":"Modeling Learner Profiles using Ontologies and Machine Learning","authors":"Samia Bousalem, Fouzia Benchikha, Massinissa Chelghoum","doi":"10.1109/NTIC55069.2022.10100497","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100497","url":null,"abstract":"In recent years, E-learning technologies have altered the way we teach and learn, making it an intriguing research topic for enhancing education. A key component of these systems is the ability to tailor the learning experience to the needs of the individual student. According to researches, modeling student profiles with an ontology is quite relevant. However, the ontology must consider every aspect of learner representation. Therefore, there is an urgent need for new comprehensive information to improve the learner profile. In this paper, we propose a semantic approach to define an ontology of learner profiles. In addition, a learning style prediction system based on machine learning techniques is developed. Empirical results show a promising gain in performance for learning style prediction systems.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126767828","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100505
Randa Nachet, T. B. Stambouli
Convolutional Neural Networks (CNNs) have shown good performance in the domain of face recognition due to their capability of extracting discriminative features. In this paper, we present a face recognition system where a Multi-Task Convolutional Neural Network (MTCNN) is employed for face detection and preprocessing. Afterwards, we use the proposed model of CNN with optimization and a softmax function as a classifier for recognition. Experiments have been carried out on the ORL face database, which consists of 400 images for 40 classes. The results of the implementation illustrate that our model has achieved better performance compared to most of the state-of-the-art models, with an accuracy rate of 97.50%.
{"title":"Improved Face Recognition Rate Using Convolutional Neural Networks","authors":"Randa Nachet, T. B. Stambouli","doi":"10.1109/NTIC55069.2022.10100505","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100505","url":null,"abstract":"Convolutional Neural Networks (CNNs) have shown good performance in the domain of face recognition due to their capability of extracting discriminative features. In this paper, we present a face recognition system where a Multi-Task Convolutional Neural Network (MTCNN) is employed for face detection and preprocessing. Afterwards, we use the proposed model of CNN with optimization and a softmax function as a classifier for recognition. Experiments have been carried out on the ORL face database, which consists of 400 images for 40 classes. The results of the implementation illustrate that our model has achieved better performance compared to most of the state-of-the-art models, with an accuracy rate of 97.50%.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126196361","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100397
Sarra Babahenini, F. Charif, A. Taleb-Ahmed
One of the numerous challenges of modern image processing is image registration. Information from many images often emerges in slightly different forms and is highly compatible. Spatial alignment is crucial to merge essential and valuable information from several images properly. The term "registration" describes this procedure. Find a transformation that results in a model that closely resembles the reference image [1].Mainly, this work is concerned with implementing two optimization algorithms: the Flower Pollination Algorithm (FPA) and the Butterfly Optimization Algorithm (BOA). To measure the efficacy of these methods, we compare the transformed image to the original by computing the mutual information between the two. The effectiveness of these methods was assessed using SSIM, EQM, and MI measures. Results from the experiments indicate that the BOA outperforms the FPA.
{"title":"Multimodal Medical Images using Rigid Iconic Registration based on Flower Pollination Algorithm and Butterfly Optimization Algorithm","authors":"Sarra Babahenini, F. Charif, A. Taleb-Ahmed","doi":"10.1109/NTIC55069.2022.10100397","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100397","url":null,"abstract":"One of the numerous challenges of modern image processing is image registration. Information from many images often emerges in slightly different forms and is highly compatible. Spatial alignment is crucial to merge essential and valuable information from several images properly. The term \"registration\" describes this procedure. Find a transformation that results in a model that closely resembles the reference image [1].Mainly, this work is concerned with implementing two optimization algorithms: the Flower Pollination Algorithm (FPA) and the Butterfly Optimization Algorithm (BOA). To measure the efficacy of these methods, we compare the transformed image to the original by computing the mutual information between the two. The effectiveness of these methods was assessed using SSIM, EQM, and MI measures. Results from the experiments indicate that the BOA outperforms the FPA.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129474797","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100585
Mohamde Amine Daoud, Abdelkader Ouared, Y. Dahmani, Sabrina Ammar
Intrusion Detection Systems (IDS) are becoming increasingly important to provide a certain level of safety in a variety of complex environments. An IDS’s proper operation is dependent on the quality of the subjective measurements that influence its quality. All steps of the IDS development life cycle must be covered to increase quality. The design phase of such a system may take long enough to show its evolution. Furthermore, each provided IDS model has a level of precision that is frequently related to the state of the system that needs to be examined. To avoid this problem, it is necessary to guide the designer in selecting suitable models and tests. In light of this, a dedicated framework has been proposed to recommend IDS instance configurations. This scope combines clustering and classification techniques to produce a resilient instance IDS analysis that ensures good independent performance from system variations. The study’s findings demonstrate that combining approaches resulted in consistent performance and high prediction accuracy.
{"title":"A Novel Pipeline to Recommend Intrusion Detection Systems Configurations","authors":"Mohamde Amine Daoud, Abdelkader Ouared, Y. Dahmani, Sabrina Ammar","doi":"10.1109/NTIC55069.2022.10100585","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100585","url":null,"abstract":"Intrusion Detection Systems (IDS) are becoming increasingly important to provide a certain level of safety in a variety of complex environments. An IDS’s proper operation is dependent on the quality of the subjective measurements that influence its quality. All steps of the IDS development life cycle must be covered to increase quality. The design phase of such a system may take long enough to show its evolution. Furthermore, each provided IDS model has a level of precision that is frequently related to the state of the system that needs to be examined. To avoid this problem, it is necessary to guide the designer in selecting suitable models and tests. In light of this, a dedicated framework has been proposed to recommend IDS instance configurations. This scope combines clustering and classification techniques to produce a resilient instance IDS analysis that ensures good independent performance from system variations. The study’s findings demonstrate that combining approaches resulted in consistent performance and high prediction accuracy.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123049732","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100455
Abderrahmane Herbadji, N. Guermat, Z. Akhtar
For developing automatic and accurate system for human recognition, deep learning is now progressively becoming common in real-world biometrics applications. Fingerprint is one of the most important discriminative biometric characteristic due to its high reliability and uniqueness properties, which has led to a widespread use by law enforcement, forensic as well as in mobile devices user authentication. Contactless fingerprint recognition has achieved rapid development in recent years thanks to more hygienic and ubiquitous personal identification techniques. In this paper, we present deep neural networks (DNNs) based solutions for contactless fingerprint identification. More specifically, we show how existing DNNs can be deployed as a feature extractor for contactless fingerprint. Experimental analyses on publically available dataset with 336 subjects demonstrate the effectiveness of DNNs-based feature extractors. Moreover, experimental results illustrate best recognition performance in comparison with state-of-the-art texture descriptors.
{"title":"Deep neural networks based contactless fingerprint recognition","authors":"Abderrahmane Herbadji, N. Guermat, Z. Akhtar","doi":"10.1109/NTIC55069.2022.10100455","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100455","url":null,"abstract":"For developing automatic and accurate system for human recognition, deep learning is now progressively becoming common in real-world biometrics applications. Fingerprint is one of the most important discriminative biometric characteristic due to its high reliability and uniqueness properties, which has led to a widespread use by law enforcement, forensic as well as in mobile devices user authentication. Contactless fingerprint recognition has achieved rapid development in recent years thanks to more hygienic and ubiquitous personal identification techniques. In this paper, we present deep neural networks (DNNs) based solutions for contactless fingerprint identification. More specifically, we show how existing DNNs can be deployed as a feature extractor for contactless fingerprint. Experimental analyses on publically available dataset with 336 subjects demonstrate the effectiveness of DNNs-based feature extractors. Moreover, experimental results illustrate best recognition performance in comparison with state-of-the-art texture descriptors.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130695886","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100492
Abdelkader Kaddour, Nassim Zellal, Lamri Sayad
The classification problem has been widely studied in data mining, machine learning, and information retrieval communities with applications in several domains, such as target marketing, medical diagnosis, newsgroup filtering, and document organization. In this work, we take up the challenge of improving Text Classification (TC) using Text Summarizing (TS).
{"title":"Improving text classification using text summarization","authors":"Abdelkader Kaddour, Nassim Zellal, Lamri Sayad","doi":"10.1109/NTIC55069.2022.10100492","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100492","url":null,"abstract":"The classification problem has been widely studied in data mining, machine learning, and information retrieval communities with applications in several domains, such as target marketing, medical diagnosis, newsgroup filtering, and document organization. In this work, we take up the challenge of improving Text Classification (TC) using Text Summarizing (TS).","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132765528","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100430
Hafida Tiaiba, L. Sabri, A. Chibani, O. Kazar
The Classification of medical reports is a crucial challenge, as they are usually presented in plain text, have a particular technical vocabulary, and are almost always unstructured. Document classification aims to assign the most appropriate label to a given document. Furthermore, among the significant issues of medical document classification is text representation in a numerical format. So, in this paper, we use artificial intelligence, proposing a model of multi-layer artificial neural networks for multi-label Classification. The transformation to numerical values of the medical documents relies on four encoding modes: Term Frequency (TF), Frequency-Inverse document frequency (TF-IDF), Bag-of- Words (BOW), and Document Term Matrix (DTM) models; in this study, we compared the four types of vectorizations. Experimental results demonstrated that the best results for our proposed neural network architecture for both models denoted Simple Neural Network (SNN) and Vocabulary Neural Network (VNN). We have used the local vocabulary of 7,400 documents in the SNN model; regarding the VNN model, we use the global terminology (Ohsumed_20000). The suggested models (VNN and SNN) performed well in classifying all four representations. Furthermore, the SNN results outperform the VNN findings. The accuracy of TF is 70.32 in time 3 with an epoch number of 64. For BOW, 68.16 is the accuracy reached with an epoch number 16 in time 1. Likewise, the accuracy of DTM with 32 epochs and in time three is 70.65, whereas the 71.08% value is the accuracy achieved by TF-IDF with 16 epochs in time 1, representing the best results obtained by SNN model.
{"title":"Artificial Neural Network for Multi-label Medical Text Classification","authors":"Hafida Tiaiba, L. Sabri, A. Chibani, O. Kazar","doi":"10.1109/NTIC55069.2022.10100430","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100430","url":null,"abstract":"The Classification of medical reports is a crucial challenge, as they are usually presented in plain text, have a particular technical vocabulary, and are almost always unstructured. Document classification aims to assign the most appropriate label to a given document. Furthermore, among the significant issues of medical document classification is text representation in a numerical format. So, in this paper, we use artificial intelligence, proposing a model of multi-layer artificial neural networks for multi-label Classification. The transformation to numerical values of the medical documents relies on four encoding modes: Term Frequency (TF), Frequency-Inverse document frequency (TF-IDF), Bag-of- Words (BOW), and Document Term Matrix (DTM) models; in this study, we compared the four types of vectorizations. Experimental results demonstrated that the best results for our proposed neural network architecture for both models denoted Simple Neural Network (SNN) and Vocabulary Neural Network (VNN). We have used the local vocabulary of 7,400 documents in the SNN model; regarding the VNN model, we use the global terminology (Ohsumed_20000). The suggested models (VNN and SNN) performed well in classifying all four representations. Furthermore, the SNN results outperform the VNN findings. The accuracy of TF is 70.32 in time 3 with an epoch number of 64. For BOW, 68.16 is the accuracy reached with an epoch number 16 in time 1. Likewise, the accuracy of DTM with 32 epochs and in time three is 70.65, whereas the 71.08% value is the accuracy achieved by TF-IDF with 16 epochs in time 1, representing the best results obtained by SNN model.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133302285","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100562
Nour Elimane Elbey, Soheyb Ayad, Bilal Benhaya
5G networks are capable of supporting a wide range of applications with different requirements, which brings several use cases for mobile networks and increases user demands. The advancement of 5G is dependent on new technologies such as Software Defined Networks (SDN), Network Function Virtualization (NFV), and Service Function Chain (SFC). SDN enables the separation of control and data planes. NFV decouples network functions from hardware using virtualization. SFC is a popular service paradigm that has been proposed to derive maximum benefits from both NFV and SDN in 5G networks. The infrastructure of 5G networks brings a change in the network management approaches for deploying network services by allocating resources and determining optimal forwarding paths. The existing deployment methods have some shortcomings that require complete knowledge of the system. For that, Reinforcement Learning (RL) and Deep Reinforcement Learning (DRL), which have demonstrated success in solving complex control and decision-making problems by allowing network entities to learn, build knowledge, and make optimal decisions separately, are used to deploy network services dynamically, which has inspired many researchers to start developing new techniques by combining machine learning approaches to solve specific networking problems. This paper reviews RL and DRL techniques that have been studied and implemented in order to deploy SFC in 5G infrastructure networks, by providing a basic description of concepts and a clear problems explication that helps new searchers invest their effort in implementing new approaches and improving existing ones.
{"title":"Review on Reinforcement Learning-based approaches for Service Function Chain deployment in 5G networks","authors":"Nour Elimane Elbey, Soheyb Ayad, Bilal Benhaya","doi":"10.1109/NTIC55069.2022.10100562","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100562","url":null,"abstract":"5G networks are capable of supporting a wide range of applications with different requirements, which brings several use cases for mobile networks and increases user demands. The advancement of 5G is dependent on new technologies such as Software Defined Networks (SDN), Network Function Virtualization (NFV), and Service Function Chain (SFC). SDN enables the separation of control and data planes. NFV decouples network functions from hardware using virtualization. SFC is a popular service paradigm that has been proposed to derive maximum benefits from both NFV and SDN in 5G networks. The infrastructure of 5G networks brings a change in the network management approaches for deploying network services by allocating resources and determining optimal forwarding paths. The existing deployment methods have some shortcomings that require complete knowledge of the system. For that, Reinforcement Learning (RL) and Deep Reinforcement Learning (DRL), which have demonstrated success in solving complex control and decision-making problems by allowing network entities to learn, build knowledge, and make optimal decisions separately, are used to deploy network services dynamically, which has inspired many researchers to start developing new techniques by combining machine learning approaches to solve specific networking problems. This paper reviews RL and DRL techniques that have been studied and implemented in order to deploy SFC in 5G infrastructure networks, by providing a basic description of concepts and a clear problems explication that helps new searchers invest their effort in implementing new approaches and improving existing ones.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116100555","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}
Pub Date : 2022-12-21DOI: 10.1109/NTIC55069.2022.10100518
Sif Eddine Boudjellal, Abdelwahhab Boudjelal, N. Boukezzoula
Breast cancer threatens the public health as it is among the leading causes of women death due to unawareness and diagnosis at the late stages. The detection of this cancer in its early stage is decisive to decrease mortality rates . Deep learning techniques are effective in analysis of medical images and achieve high performance in detecting the abnormal features, and classify them. Therefore, these methods are becoming increasingly popular in breast cancer diagnosis. Convolutional Neural Networks (CNNs) are commonly used for medical image analysis, but Vision transformers (ViTs ) are becoming more popular due to their excellent performance. However, ViTs still fall behind state-of-the-art convolutional networks. To overcome these limitations, many researchers have proposed a new approach that combines the advantages of CNNs and Transformers. This new approach overcomes the limitations of each by extracting low-level features, strengthening locality, and establishing long-range dependencies. In this study, the Hybrid Conv-Transformer approach was used to extract features from the BreakHis dataset of histopathological images. Coatnet and ConvMixer models were then used to classify the images into two binary classification based on both magnification-dependent and magnification-independent categories. The findings indicated that the suggested models exceeded prior models and recent deep learning techniques on the BreakHis dataset.
{"title":"Hybrid Convolution-Transformer models for breast cancer classification using histopathological images","authors":"Sif Eddine Boudjellal, Abdelwahhab Boudjelal, N. Boukezzoula","doi":"10.1109/NTIC55069.2022.10100518","DOIUrl":"https://doi.org/10.1109/NTIC55069.2022.10100518","url":null,"abstract":"Breast cancer threatens the public health as it is among the leading causes of women death due to unawareness and diagnosis at the late stages. The detection of this cancer in its early stage is decisive to decrease mortality rates . Deep learning techniques are effective in analysis of medical images and achieve high performance in detecting the abnormal features, and classify them. Therefore, these methods are becoming increasingly popular in breast cancer diagnosis. Convolutional Neural Networks (CNNs) are commonly used for medical image analysis, but Vision transformers (ViTs ) are becoming more popular due to their excellent performance. However, ViTs still fall behind state-of-the-art convolutional networks. To overcome these limitations, many researchers have proposed a new approach that combines the advantages of CNNs and Transformers. This new approach overcomes the limitations of each by extracting low-level features, strengthening locality, and establishing long-range dependencies. In this study, the Hybrid Conv-Transformer approach was used to extract features from the BreakHis dataset of histopathological images. Coatnet and ConvMixer models were then used to classify the images into two binary classification based on both magnification-dependent and magnification-independent categories. The findings indicated that the suggested models exceeded prior models and recent deep learning techniques on the BreakHis dataset.","PeriodicalId":403927,"journal":{"name":"2022 2nd International Conference on New Technologies of Information and Communication (NTIC)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122120329","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}
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：