Pub Date : 2023-11-03DOI: 10.4015/s101623722350028x
Samla Salim, R. Sarath
Cancer is the second cause of mortality worldwide and it has been identified as a perilous disease. Among all types of cancers, Breast Cancer (BC) is a substantial research subject in the medical imaging area, because it is a serious disease and primary reason for death in women. Proper diagnosis helps patients to get adequate treatment, enhancing the probability of surviving. Because of the poor contrast and unclear structure of tumor cells in the images, automatic segmenting of breast tumors remains a difficult task. However, the identification and interpretation of breast lesions is challenging even for expert radiologists. To address these limitations, an efficient mechanism for BC detection and classification using histopathological images is proposed, which employs a DenseNet-based Chronological Circle Inspired Optimization Algorithm (CCIOA). Deep Learning (DL) approaches are used in the suggested BC classification scheme to precisely segment and identify the BC. The segmentation is done using ResuNet++, and an efficient optimization method called Invasive Water Ebola Optimization (IWEO) is used to fine-tune the DL network’s parameters. Furthermore, DenseNet is utilized for BC detection, while CCIOA is used for DenseNet training. The CCIOA-DenseNet is evaluated using the metrics of accuracy, True Positive Rate (TPR), and True Negative Rate (TNR). Experiment results show that the CCIOA-DenseNet attained better accuracy of 0.971, TPR of 0.966, and TNR of 0.954.
{"title":"BREAST CANCER DETECTION AND CLASSIFICATION USING HISTOPATHOLOGICAL IMAGES BASED ON OPTIMIZATION-ENABLED DEEP LEARNING","authors":"Samla Salim, R. Sarath","doi":"10.4015/s101623722350028x","DOIUrl":"https://doi.org/10.4015/s101623722350028x","url":null,"abstract":"Cancer is the second cause of mortality worldwide and it has been identified as a perilous disease. Among all types of cancers, Breast Cancer (BC) is a substantial research subject in the medical imaging area, because it is a serious disease and primary reason for death in women. Proper diagnosis helps patients to get adequate treatment, enhancing the probability of surviving. Because of the poor contrast and unclear structure of tumor cells in the images, automatic segmenting of breast tumors remains a difficult task. However, the identification and interpretation of breast lesions is challenging even for expert radiologists. To address these limitations, an efficient mechanism for BC detection and classification using histopathological images is proposed, which employs a DenseNet-based Chronological Circle Inspired Optimization Algorithm (CCIOA). Deep Learning (DL) approaches are used in the suggested BC classification scheme to precisely segment and identify the BC. The segmentation is done using ResuNet++, and an efficient optimization method called Invasive Water Ebola Optimization (IWEO) is used to fine-tune the DL network’s parameters. Furthermore, DenseNet is utilized for BC detection, while CCIOA is used for DenseNet training. The CCIOA-DenseNet is evaluated using the metrics of accuracy, True Positive Rate (TPR), and True Negative Rate (TNR). Experiment results show that the CCIOA-DenseNet attained better accuracy of 0.971, TPR of 0.966, and TNR of 0.954.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"28 13","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135873589","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}
Electroencephalography (EEG) is a common clinical method of recording the electrical activity of the brain. EEG can record High-Frequency Oscillations ([Formula: see text]80 HZ), which carry appropriate information regarding Epilepsy. High-Frequency Oscillations (HFO) serve as a potential biomarker for Epileptogenesis. EEG signals are often prone to artifact corruptions, which mislead the clinicians by the incorrect signal interpretations. Therefore, automatic artifact removal approach is a key phase in all the Brain-Computer Interface (BCI) applications. In this work, the automatic artifact identification and removal strategy without consuming any supplementary reference channel using two different approaches is developed and discussed. A proficient novel Modified Online Bi-Conjugate Gradient-based Independent Component Analysis (MOBICA) is developed. An efficient threshold-based peak detection and removal strategy, Sparsity-based Artifact Removal Technique (SART) is constructed, where Principle Component Analysis (PCA) is replaced with Singular Value Decomposition (SVD) of the K-SVD algorithm. Both the proposed models are evaluated on two different datasets like CHB-MIT and SRM scalp data recordings. Both the MOBICA and SART algorithms removed the artifactual component parting the intact EEG source component. Finally, the performance of the proposed agenda is compared with the conventional approaches. Our MOBICA and SART algorithms remove the artifactual component parting the intact EEG source component. Empirical results of SART on CHB-MIT and SRM databases of 52 EEG recordings outperform MOBICA maintaining least Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and high Signal to Artifact Ratio (SAR), Mutual Information (MI), and Correlation Coefficient (CC). The proposed strategy vows to be a promising solution for artifact removal in the clinical use of EEG signals and in BCI applications.
{"title":"AUTOMATIC ELECTROENCEPHALOGRAPHIC SOURCE SEPARATION STRATEGIES FOR SEIZURE PREDICTION APPLICATION","authors":"Banu Priya Prathaban, Subash Rajendran, Ramachandran Balasubramanian","doi":"10.4015/s1016237223500321","DOIUrl":"https://doi.org/10.4015/s1016237223500321","url":null,"abstract":"Electroencephalography (EEG) is a common clinical method of recording the electrical activity of the brain. EEG can record High-Frequency Oscillations ([Formula: see text]80 HZ), which carry appropriate information regarding Epilepsy. High-Frequency Oscillations (HFO) serve as a potential biomarker for Epileptogenesis. EEG signals are often prone to artifact corruptions, which mislead the clinicians by the incorrect signal interpretations. Therefore, automatic artifact removal approach is a key phase in all the Brain-Computer Interface (BCI) applications. In this work, the automatic artifact identification and removal strategy without consuming any supplementary reference channel using two different approaches is developed and discussed. A proficient novel Modified Online Bi-Conjugate Gradient-based Independent Component Analysis (MOBICA) is developed. An efficient threshold-based peak detection and removal strategy, Sparsity-based Artifact Removal Technique (SART) is constructed, where Principle Component Analysis (PCA) is replaced with Singular Value Decomposition (SVD) of the K-SVD algorithm. Both the proposed models are evaluated on two different datasets like CHB-MIT and SRM scalp data recordings. Both the MOBICA and SART algorithms removed the artifactual component parting the intact EEG source component. Finally, the performance of the proposed agenda is compared with the conventional approaches. Our MOBICA and SART algorithms remove the artifactual component parting the intact EEG source component. Empirical results of SART on CHB-MIT and SRM databases of 52 EEG recordings outperform MOBICA maintaining least Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and high Signal to Artifact Ratio (SAR), Mutual Information (MI), and Correlation Coefficient (CC). The proposed strategy vows to be a promising solution for artifact removal in the clinical use of EEG signals and in BCI applications.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"33 S1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135973483","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 : 2023-10-20DOI: 10.4015/s1016237223500308
Nithyakalyani Krishnan, P. Muthu
Chronic Venous Insufficiency (CVI) is a venous incompetence condition that leads to improper blood circulation from the lower limbs towards the heart. This occurs as a result of blood pooling in the veins of the leg, resulting in twisted, dilated, and tortuous veins. Aging, obesity, prolonged standing or sitting, and lack of mobility are all important causes of the occurrence of this chronic disease. The cost of CVI diagnosis and treatment is extremely high. Infrared thermographic image analysis is used for early detection and reduces the cost of diagnosis. Deep learning (DL) techniques play an important role in early prediction and may aid clinicians in diagnosing CVI. An automated classification model will assist the physician in making a precise diagnosis of the abnormal vein and treating the patient according to the severity of the condition. There is a need for a model that can perform successful classification without the need for pre-processing when compared to the traditional machine learning (ML) methods that depend on ideal manual feature extraction to achieve optimal outcomes. In this research, we recommend the customized DenseNet-121 architecture for CVI detection and compare it with other advanced DL models to determine its efficacy. DenseNet-121 and other pre-trained convolutional neural network models, including EfficientNetB0 and Inception_v3, were trained using a transfer learning strategy. The experimental findings indicate that the proposed modified DenseNet-121 model outperformed other classical methods. The reported results provide evidence of the robustness of the suggested method in addition to the high accuracy that it possessed, as shown by the overall testing accuracy of 97.4%. Thus, this study can be considered as a non-invasive and cost-effective approach for diagnosing chronic venous insufficiency condition in lower extremity.
{"title":"DETECTION OF CHRONIC VENOUS INSUFFICIENCY CONDITION USING TRANSFER LEARNING WITH CONVOLUTIONAL NEURAL NETWORKS BASED ON THERMAL IMAGES","authors":"Nithyakalyani Krishnan, P. Muthu","doi":"10.4015/s1016237223500308","DOIUrl":"https://doi.org/10.4015/s1016237223500308","url":null,"abstract":"Chronic Venous Insufficiency (CVI) is a venous incompetence condition that leads to improper blood circulation from the lower limbs towards the heart. This occurs as a result of blood pooling in the veins of the leg, resulting in twisted, dilated, and tortuous veins. Aging, obesity, prolonged standing or sitting, and lack of mobility are all important causes of the occurrence of this chronic disease. The cost of CVI diagnosis and treatment is extremely high. Infrared thermographic image analysis is used for early detection and reduces the cost of diagnosis. Deep learning (DL) techniques play an important role in early prediction and may aid clinicians in diagnosing CVI. An automated classification model will assist the physician in making a precise diagnosis of the abnormal vein and treating the patient according to the severity of the condition. There is a need for a model that can perform successful classification without the need for pre-processing when compared to the traditional machine learning (ML) methods that depend on ideal manual feature extraction to achieve optimal outcomes. In this research, we recommend the customized DenseNet-121 architecture for CVI detection and compare it with other advanced DL models to determine its efficacy. DenseNet-121 and other pre-trained convolutional neural network models, including EfficientNetB0 and Inception_v3, were trained using a transfer learning strategy. The experimental findings indicate that the proposed modified DenseNet-121 model outperformed other classical methods. The reported results provide evidence of the robustness of the suggested method in addition to the high accuracy that it possessed, as shown by the overall testing accuracy of 97.4%. Thus, this study can be considered as a non-invasive and cost-effective approach for diagnosing chronic venous insufficiency condition in lower extremity.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135618398","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 : 2023-10-01DOI: 10.4015/s1016237223500278
Hossein Kargar Jahromi, Morteza Alizadeh, Arian Ehterami, Ahmad Vaez, Danial Cheraghali, Leila Chegini, Nariman Rezaei Kolarijani, Majid Salehi
Bone defects affect many people and impose expenses of costly treatment with possible complications. This study aims to investigate a novel Poly ([Formula: see text]-caprolactone)/Poly (L-lactic) acid/Gelatin nanofiber [PCL/PLA/GNF] scaffold containing 5% resveratrol (Resv) which was manufactured via thermally induced phase separation technique (TIPS), and its applicability for bone defect treatment. Gelatin nanofiber (GNF) was synthesized via the electrospinning method and mixed with PCL/PLA solution and then 5% resveratrol was added to fabricate a 3D scaffold via the TIPS technique. The prepared scaffolds were evaluated regarding their porosity, morphology, contact angle, degradation properties, biomechanical, blood compatibility, and cell viability via MTT assay. The scaffolds were further investigated by implantation in a rat femur defect model. PCL/PLA/GNF with 5% Resv showed a cancellated structure with irregular-shaped pores. The mean pore size was estimated to be 160 [Formula: see text]m and the porosity was 80.56 ± 2.68%. The contact angle of the fabricated scaffold was 95.4 ± 3.4, which determines the hydrophobic nature of the scaffold. Increased cell viability in scaffolds was observed by adding resveratrol. Twelve weeks after the implantation of the scaffold into the bone defect, the defects filled with PCL/PLA/GNF-resveratrol contained scaffold were remarkably better than PCL/PLA/GNF and negative control group (89.23 ± 6.34% in 12 weeks), and the difference was significant (p ¡ 0.05). In conclusion, the PCL/PLA/GNF scaffold containing 5% of resveratrol demonstrated adequate mechanical and physical properties. There is possible applicability of PCL/PLA/GNF scaffold containing 5% of resveratrol for surgical treatment of bone defects.
{"title":"EVALUATION OF THE EFFECT OF POLY (𝜀-CAPROLACTONE)/POLY (L-LACTIC) ACID/GELATIN NANOFIBER 3D SCAFFOLD CONTAINING RESVERATROL ON BONE REGENERATION","authors":"Hossein Kargar Jahromi, Morteza Alizadeh, Arian Ehterami, Ahmad Vaez, Danial Cheraghali, Leila Chegini, Nariman Rezaei Kolarijani, Majid Salehi","doi":"10.4015/s1016237223500278","DOIUrl":"https://doi.org/10.4015/s1016237223500278","url":null,"abstract":"Bone defects affect many people and impose expenses of costly treatment with possible complications. This study aims to investigate a novel Poly ([Formula: see text]-caprolactone)/Poly (L-lactic) acid/Gelatin nanofiber [PCL/PLA/GNF] scaffold containing 5% resveratrol (Resv) which was manufactured via thermally induced phase separation technique (TIPS), and its applicability for bone defect treatment. Gelatin nanofiber (GNF) was synthesized via the electrospinning method and mixed with PCL/PLA solution and then 5% resveratrol was added to fabricate a 3D scaffold via the TIPS technique. The prepared scaffolds were evaluated regarding their porosity, morphology, contact angle, degradation properties, biomechanical, blood compatibility, and cell viability via MTT assay. The scaffolds were further investigated by implantation in a rat femur defect model. PCL/PLA/GNF with 5% Resv showed a cancellated structure with irregular-shaped pores. The mean pore size was estimated to be 160 [Formula: see text]m and the porosity was 80.56 ± 2.68%. The contact angle of the fabricated scaffold was 95.4 ± 3.4, which determines the hydrophobic nature of the scaffold. Increased cell viability in scaffolds was observed by adding resveratrol. Twelve weeks after the implantation of the scaffold into the bone defect, the defects filled with PCL/PLA/GNF-resveratrol contained scaffold were remarkably better than PCL/PLA/GNF and negative control group (89.23 ± 6.34% in 12 weeks), and the difference was significant (p ¡ 0.05). In conclusion, the PCL/PLA/GNF scaffold containing 5% of resveratrol demonstrated adequate mechanical and physical properties. There is possible applicability of PCL/PLA/GNF scaffold containing 5% of resveratrol for surgical treatment of bone defects.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136054903","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 : 2023-09-30DOI: 10.4015/s1016237223500266
Jothiraj Selvaraj, A. K. Jayanthy
Colorectal cancer (CRC), ranking third most prevalent cancer type, can be diagnosed with the detection of polyps in the colon and rectum through endoscopic procedures facilitating prompt treatment. During visualization of gastrointestinal tract by the physician, there is high probability of miss rates and reviewing of the images is laborious. Automatic segmentation and detection are enabled with the convolutional neural networks (CNN). We segmented the polyps from the wireless capsule endoscopy images of Kvasir dataset using various CNN models. We have presented nine optimizers for each architecture and evaluated the performance parameters. The optimizers were graded based on the performance metrics in order to provide an insight for the researchers on the selection of optimizer and architecture. On comparison of the performance metrics of the pretrained and U-net-based architecture, the Adaptive Moment Estimation (ADAM) and Root Mean Squared Propagation (RMSPROP) optimizers received the highest score of 43 in the ranking, DiffGrad and Nesterov-accelerated Adaptive Moment Estimation (NADAM) ranked second with the score of 13, the Adaptive Delta (ADADELTA) ranked third with a score of 2, whereas Stochastic Gradient Descent (SGD), Adaptive Gradient Descent (ADAGRAD), and Adaptive Max (ADAMAX) optimizers performed least in the evaluation. Based on the deep learning application, the optimizer employed varies by considering computational speed, memory and computational time. This preliminary research provides the necessary key information for consideration in the development of an architecture with utilization of an optimizer.
{"title":"AUTOMATIC POLYP SEMANTIC SEGMENTATION USING WIRELESS CAPSULE ENDOSCOPY IMAGES WITH VARIOUS CONVOLUTIONAL NEURAL NETWORK AND OPTIMIZATION TECHNIQUES: A COMPARISON AND PERFORMANCE EVALUATION","authors":"Jothiraj Selvaraj, A. K. Jayanthy","doi":"10.4015/s1016237223500266","DOIUrl":"https://doi.org/10.4015/s1016237223500266","url":null,"abstract":"Colorectal cancer (CRC), ranking third most prevalent cancer type, can be diagnosed with the detection of polyps in the colon and rectum through endoscopic procedures facilitating prompt treatment. During visualization of gastrointestinal tract by the physician, there is high probability of miss rates and reviewing of the images is laborious. Automatic segmentation and detection are enabled with the convolutional neural networks (CNN). We segmented the polyps from the wireless capsule endoscopy images of Kvasir dataset using various CNN models. We have presented nine optimizers for each architecture and evaluated the performance parameters. The optimizers were graded based on the performance metrics in order to provide an insight for the researchers on the selection of optimizer and architecture. On comparison of the performance metrics of the pretrained and U-net-based architecture, the Adaptive Moment Estimation (ADAM) and Root Mean Squared Propagation (RMSPROP) optimizers received the highest score of 43 in the ranking, DiffGrad and Nesterov-accelerated Adaptive Moment Estimation (NADAM) ranked second with the score of 13, the Adaptive Delta (ADADELTA) ranked third with a score of 2, whereas Stochastic Gradient Descent (SGD), Adaptive Gradient Descent (ADAGRAD), and Adaptive Max (ADAMAX) optimizers performed least in the evaluation. Based on the deep learning application, the optimizer employed varies by considering computational speed, memory and computational time. This preliminary research provides the necessary key information for consideration in the development of an architecture with utilization of an optimizer.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"105 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136342558","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 : 2023-09-27DOI: 10.4015/s1016237223500199
Roghayyeh Arvanaghi, Saeed Meshgini
Background and Objective: Regarding the Coronavirus disease-2019 (COVID-19) pandemic in past years and using medical images to detect it, the image processing of the lungs and enhancement of its quality are some of the challenges in the medical image processing field. As it sounds from previous studies, the lung image processing has been raised in the other lung diseases such as lung cancer, too. Thus, the accurate classifying between normal lung image and abnormal is a challenge to aid physicians. Methods: In this paper, we have proposed an image fusion technique to increase the accuracy of classifier. In this technique, some signal preprocessing tools like discrete wavelet transform (DWT), wavelet scattering transform (WST), and image fusion by using DWT are employed to enhance ordinary convolutional neural network (CNN) classifier accuracy. Results: Unlike other studies, in this paper, different aspects of an image are fused with itself to emphasize its information which may be neglected in a total assessment of the image. We have achieved 89.8% accuracy for very simple structure of CNN classifier without using proposed fusion, and when we used proposed methods, the classifier accuracy increased to 91.8%. Conclusions: This study reveals using efficient preprocessing and presenting input images which lead to decrease the complications of deep learning classifier, and increase its accuracy overall.
{"title":"EVALUATION OF THE EFFECTS OF LUNGS CHEST X-RAY IMAGE FUSION WITH ITS WAVELET SCATTERING TRANSFORM COEFFICIENTS ON THE CONVENTIONAL NEURAL NETWORK CLASSIFIER ACCURACY DURING THE COVID-19 DISEASE","authors":"Roghayyeh Arvanaghi, Saeed Meshgini","doi":"10.4015/s1016237223500199","DOIUrl":"https://doi.org/10.4015/s1016237223500199","url":null,"abstract":"Background and Objective: Regarding the Coronavirus disease-2019 (COVID-19) pandemic in past years and using medical images to detect it, the image processing of the lungs and enhancement of its quality are some of the challenges in the medical image processing field. As it sounds from previous studies, the lung image processing has been raised in the other lung diseases such as lung cancer, too. Thus, the accurate classifying between normal lung image and abnormal is a challenge to aid physicians. Methods: In this paper, we have proposed an image fusion technique to increase the accuracy of classifier. In this technique, some signal preprocessing tools like discrete wavelet transform (DWT), wavelet scattering transform (WST), and image fusion by using DWT are employed to enhance ordinary convolutional neural network (CNN) classifier accuracy. Results: Unlike other studies, in this paper, different aspects of an image are fused with itself to emphasize its information which may be neglected in a total assessment of the image. We have achieved 89.8% accuracy for very simple structure of CNN classifier without using proposed fusion, and when we used proposed methods, the classifier accuracy increased to 91.8%. Conclusions: This study reveals using efficient preprocessing and presenting input images which lead to decrease the complications of deep learning classifier, and increase its accuracy overall.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"40 8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135538455","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 : 2023-08-31DOI: 10.4015/s1016237223500254
R. Shalini, V. Gopi
An accurate assessment of the morphological characteristics of the Optic Disc (OD) is essential for the diagnosis of various retinal disorders. It is necessary to segment the OD precisely to detect structural OD changes associated with visual field loss. Although deep learning models are effective for this task, they require extensive labeled datasets, which can be time-consuming and costly. Furthermore, fundus images have multi-scale features, making segmentation challenging. In this study, we present a semi-supervised and transfer learning approach for OD segmentation. Our approach utilizes an im-proved Dilated Convolutional AutoEncoder (DCAE) and a pre-trained modified U-Net to segment the OD. The DCAE seg-ments the OD using feature similarity from unlabeled images in the Messidor dataset and saves the learned weights. Trans-fer learning is then applied to reuse the model weights in the U-Net, accelerating training on small datasets such as Drions-DB and Drishti-GS. The network architecture was modified by increasing the layers from 8 to 128 and halving the feature map length and width. To address the multi-scale challenge without inflating the model parameters, we introduce the Dilated Hierarchical Feature Extraction Module (DHFEM), a convolutional module capable of achieving multi-scale feature extraction without increasing model parameters. Additionally, DHFEM incorporates convolutional layers with varying recep-tive fields, further enhancing the network ability to extract features across multiple scales. Our OD segmentation method outperforms existing algorithms with reduced parameter quantities of 0.4 M. The mean Intersection over Union (mIoU) values are 0.9383 and 0.9629 and inference times of 45 ms and 40 ms for the Drions-DB and Drishti-GS datasets, respectively.
{"title":"DCAE-UNET: IMPROVED OPTIC DISC SEGMENTATION MODEL USING SEMI-SUPERVISED DEEP DILATED CONVOLUTION AUTOENCODER-BASED MODIFIED U-NET","authors":"R. Shalini, V. Gopi","doi":"10.4015/s1016237223500254","DOIUrl":"https://doi.org/10.4015/s1016237223500254","url":null,"abstract":"An accurate assessment of the morphological characteristics of the Optic Disc (OD) is essential for the diagnosis of various retinal disorders. It is necessary to segment the OD precisely to detect structural OD changes associated with visual field loss. Although deep learning models are effective for this task, they require extensive labeled datasets, which can be time-consuming and costly. Furthermore, fundus images have multi-scale features, making segmentation challenging. In this study, we present a semi-supervised and transfer learning approach for OD segmentation. Our approach utilizes an im-proved Dilated Convolutional AutoEncoder (DCAE) and a pre-trained modified U-Net to segment the OD. The DCAE seg-ments the OD using feature similarity from unlabeled images in the Messidor dataset and saves the learned weights. Trans-fer learning is then applied to reuse the model weights in the U-Net, accelerating training on small datasets such as Drions-DB and Drishti-GS. The network architecture was modified by increasing the layers from 8 to 128 and halving the feature map length and width. To address the multi-scale challenge without inflating the model parameters, we introduce the Dilated Hierarchical Feature Extraction Module (DHFEM), a convolutional module capable of achieving multi-scale feature extraction without increasing model parameters. Additionally, DHFEM incorporates convolutional layers with varying recep-tive fields, further enhancing the network ability to extract features across multiple scales. Our OD segmentation method outperforms existing algorithms with reduced parameter quantities of 0.4 M. The mean Intersection over Union (mIoU) values are 0.9383 and 0.9629 and inference times of 45 ms and 40 ms for the Drions-DB and Drishti-GS datasets, respectively.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"49 1","pages":""},"PeriodicalIF":0.9,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76796580","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 : 2023-08-31DOI: 10.4015/s1016237223500230
Chinnu Jacob, C. Gopakumar, Fathima Nazarudeen
Lung cancer is a major global health concern and a leading cause of cancer-related deaths. Accurate diagnosis and treatment of lung cancer are crucial for improving patient outcomes. Subtyping lung cancer provides essential information about its molecular characteristics, clinical behavior, and prognosis, thereby guiding treatment planning. Radiomics, a novel discipline, offers a promising approach to characterize the tumor microenvironment by extracting quantitative imaging features from medical images. Radiomics aims to comprehensively and non-invasively characterize tumors and their microenvironment, enabling the identification of tumor subtypes, prediction of therapy response, and enhancement of patient outcomes. This study evaluates the effectiveness of a Particle Swarm Optimization-Random Forest (PSO-RF) classifier for subtype categorization of lung cancer based on radiomics using computed tomography (CT) images. The study utilizes three datasets, extracting 1093 radiomic features and reducing them to 20 significant features through extra tree feature selection. Optimized parameters of the PSO-RF classifier are determined using 10-fold cross-validation and compared to traditional machine learning classifiers and reported works. Results demonstrate that the PSO-RF classifier outperforms other methods, achieving an accuracy of 92%, precision of 92.5%, recall of 92%, and [Formula: see text] 1-score of 0.92 in the Lung1 dataset. Training on Dataset 3 and validating the Lung3 dataset confirm the generalizability of the model, yielding an accuracy of 87% and an AUC of 0.91 across diverse scenarios. These findings highlight the efficacy of radiomics in identifying lung cancer subtypes and demonstrate the potential of the PSO-RF classifier. The incorporation of radiomics into clinical practice has the potential to greatly improve patient outcomes by customizing treatment approaches according to unique tumor characteristics. The demonstrated effectiveness of the PSO-RF classifier makes it a valuable resource for diagnosing and categorizing different subtypes of lung cancer.
{"title":"OPTIMIZED RADIOMICS-BASED MACHINE LEARNING APPROACH FOR LUNG CANCER SUBTYPE CLASSIFICATION","authors":"Chinnu Jacob, C. Gopakumar, Fathima Nazarudeen","doi":"10.4015/s1016237223500230","DOIUrl":"https://doi.org/10.4015/s1016237223500230","url":null,"abstract":"Lung cancer is a major global health concern and a leading cause of cancer-related deaths. Accurate diagnosis and treatment of lung cancer are crucial for improving patient outcomes. Subtyping lung cancer provides essential information about its molecular characteristics, clinical behavior, and prognosis, thereby guiding treatment planning. Radiomics, a novel discipline, offers a promising approach to characterize the tumor microenvironment by extracting quantitative imaging features from medical images. Radiomics aims to comprehensively and non-invasively characterize tumors and their microenvironment, enabling the identification of tumor subtypes, prediction of therapy response, and enhancement of patient outcomes. This study evaluates the effectiveness of a Particle Swarm Optimization-Random Forest (PSO-RF) classifier for subtype categorization of lung cancer based on radiomics using computed tomography (CT) images. The study utilizes three datasets, extracting 1093 radiomic features and reducing them to 20 significant features through extra tree feature selection. Optimized parameters of the PSO-RF classifier are determined using 10-fold cross-validation and compared to traditional machine learning classifiers and reported works. Results demonstrate that the PSO-RF classifier outperforms other methods, achieving an accuracy of 92%, precision of 92.5%, recall of 92%, and [Formula: see text] 1-score of 0.92 in the Lung1 dataset. Training on Dataset 3 and validating the Lung3 dataset confirm the generalizability of the model, yielding an accuracy of 87% and an AUC of 0.91 across diverse scenarios. These findings highlight the efficacy of radiomics in identifying lung cancer subtypes and demonstrate the potential of the PSO-RF classifier. The incorporation of radiomics into clinical practice has the potential to greatly improve patient outcomes by customizing treatment approaches according to unique tumor characteristics. The demonstrated effectiveness of the PSO-RF classifier makes it a valuable resource for diagnosing and categorizing different subtypes of lung cancer.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"43 1","pages":""},"PeriodicalIF":0.9,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89168071","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 : 2023-08-31DOI: 10.4015/s1016237223500229
S. Sowmiya, U. Snehalatha, Jayanth Murugan
Renal cysts are categorized as simple cysts and complex cysts. Simple cysts are harmless and complicated cysts are cancerous and leading to a dangerous situation. The study aims to implement a deep learning-based segmentation that uses the Renal images to segment the cyst, detecting the size of the cyst and assessing the state of cyst from the infected renal image. The automated method for segmenting renal cysts from MRI abdominal images is based on a U-net algorithm. The deep learning-based segmentation like U-net algorithm segmented the renal cyst. The characteristics of the segmented cyst were analyzed using the Statistical features extracted using GLCM algorithm. The machine learning classification is performed using the extracted GLCM features. Three machine learning classifiers such as Naïve Bayes, Hoeffding Tree and SVM are used in the proposed study. Naive Bayes and Hoeffding Tree achieved the highest accuracy of 98%. The SVM classifier achieved 96% of accuracy. This study proposed a new system to diagnose the renal cyst from MRI abdomen images. Our study focused on cyst segmentation, size detection, feature extraction and classification. The three-classification method suits best for classifying the renal cyst. Naïve Bayes and Hoeffding Tree classifier achieved the highest accuracy. The diameter of cyst size is measured using the blobs analysis method to predict the renal cyst at an earlier stage. Hence, the deep learning-based segmentation performed well in segmenting the renal cyst and the three classifiers achieved the highest accuracy, above 95%.
{"title":"RENAL CYST DETECTION IN ABDOMINAL MRI IMAGES USING DEEP LEARNING SEGMENTATION","authors":"S. Sowmiya, U. Snehalatha, Jayanth Murugan","doi":"10.4015/s1016237223500229","DOIUrl":"https://doi.org/10.4015/s1016237223500229","url":null,"abstract":"Renal cysts are categorized as simple cysts and complex cysts. Simple cysts are harmless and complicated cysts are cancerous and leading to a dangerous situation. The study aims to implement a deep learning-based segmentation that uses the Renal images to segment the cyst, detecting the size of the cyst and assessing the state of cyst from the infected renal image. The automated method for segmenting renal cysts from MRI abdominal images is based on a U-net algorithm. The deep learning-based segmentation like U-net algorithm segmented the renal cyst. The characteristics of the segmented cyst were analyzed using the Statistical features extracted using GLCM algorithm. The machine learning classification is performed using the extracted GLCM features. Three machine learning classifiers such as Naïve Bayes, Hoeffding Tree and SVM are used in the proposed study. Naive Bayes and Hoeffding Tree achieved the highest accuracy of 98%. The SVM classifier achieved 96% of accuracy. This study proposed a new system to diagnose the renal cyst from MRI abdomen images. Our study focused on cyst segmentation, size detection, feature extraction and classification. The three-classification method suits best for classifying the renal cyst. Naïve Bayes and Hoeffding Tree classifier achieved the highest accuracy. The diameter of cyst size is measured using the blobs analysis method to predict the renal cyst at an earlier stage. Hence, the deep learning-based segmentation performed well in segmenting the renal cyst and the three classifiers achieved the highest accuracy, above 95%.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"29 1","pages":""},"PeriodicalIF":0.9,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75751908","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 : 2023-08-31DOI: 10.4015/s1016237223500242
Zhuyan Lyu, Q. Xue, Shuo Yang, Meng Jiao, Kai Qi
Background: Sit-to-stand (STS) is an integral daily life activity. Handrail height significantly affects STS. However, the multifactorial influences of horizontal handrail height and width on STS have not been investigated. Purpose: The purpose of this study was to evaluate the influence of different heights and widths of horizontal handrails on the motion time, joint angles, and joint moments during STS to determine the optimal handrail height and width during STS. Methods:The study was conducted on 16 healthy young adults. Six experimental conditions were tested: high handrail large width; high handrail small width; medium handrail large width; middle handrail small width; low handrail large width; low handrail small width. The movement time, joint angle, and joint moment were analyzed and compared. Results: Different handrail heights had a significant influence on the percent of motion time in the first phase. Only handrail height significantly influenced the maximum trunk tilt angle. There was an interaction between handrail height and width for the peak hip joint moment. Conclusions: These findings indicated that people who have difficulty leaning forward will expend less effort and backward falls can be prevented when using the high handrail. The large width can be particularly helpful for patients with poor hip strength. Therefore, patients with impaired lower extremity strength can employ a high handrail with a large width to reduce the burden of performing STS transfers.
{"title":"EFFECT OF DIFFERENT HANDRAIL HEIGHTS AND WIDTHS ON KINEMATICS AND KINETICS OF SIT-TO-STAND IN HEALTHY YOUNG ADULTS","authors":"Zhuyan Lyu, Q. Xue, Shuo Yang, Meng Jiao, Kai Qi","doi":"10.4015/s1016237223500242","DOIUrl":"https://doi.org/10.4015/s1016237223500242","url":null,"abstract":"Background: Sit-to-stand (STS) is an integral daily life activity. Handrail height significantly affects STS. However, the multifactorial influences of horizontal handrail height and width on STS have not been investigated. Purpose: The purpose of this study was to evaluate the influence of different heights and widths of horizontal handrails on the motion time, joint angles, and joint moments during STS to determine the optimal handrail height and width during STS. Methods:The study was conducted on 16 healthy young adults. Six experimental conditions were tested: high handrail large width; high handrail small width; medium handrail large width; middle handrail small width; low handrail large width; low handrail small width. The movement time, joint angle, and joint moment were analyzed and compared. Results: Different handrail heights had a significant influence on the percent of motion time in the first phase. Only handrail height significantly influenced the maximum trunk tilt angle. There was an interaction between handrail height and width for the peak hip joint moment. Conclusions: These findings indicated that people who have difficulty leaning forward will expend less effort and backward falls can be prevented when using the high handrail. The large width can be particularly helpful for patients with poor hip strength. Therefore, patients with impaired lower extremity strength can employ a high handrail with a large width to reduce the burden of performing STS transfers.","PeriodicalId":8862,"journal":{"name":"Biomedical Engineering: Applications, Basis and Communications","volume":"57 1","pages":""},"PeriodicalIF":0.9,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81502223","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：