Pub Date : 2023-06-01DOI: 10.18178/joig.11.2.178-184
Daniel Kipele, K. Greyson
The presence of Poisson noise in medical X-ray images leads to degradation of the image quality. The obscured information is required for accurate diagnosis. During X-ray image acquisition process, weak light results into limited number of available photons, which leads into the Poisson noise commonly known as X-ray noise. Currently, the available X-ray noise removal methods have not yet obtained satisfying total denoising results to remove noise from the medical X-ray images. The available techniques tend to show good performance when the image model corresponds to the algorithm’s assumptions used but in general, the denoising algorithms fail to do complete denoise. X-ray image quality could be improved by increasing the X-ray dose value (beyond the maximum medically permissible dose) but the process could be lethal to patients’ health since higher X-ray energy may kill cells due to the effects of higher dose values. In this study, the hybrid model that combines the Poisson Principal Component Analysis (Poisson PCA) with the nonlocal (NL) means denoising algorithm is developed to reduce noise in images. This hybrid model for X-ray noise removal and the contrast enhancement improves the quality of X-ray images and can, thus, be used for medical diagnosis. The performance of the proposed hybrid model was observed by using the standard data and was compared with the standard Poisson algorithms.
医用x射线图像中泊松噪声的存在会导致图像质量的下降。准确诊断需要模糊的信息。在x射线图像采集过程中，弱光导致可用光子数量有限，从而导致通常称为x射线噪声的泊松噪声。目前，现有的x射线去噪方法尚未获得令人满意的全去噪效果，以去除医学x射线图像中的噪声。当图像模型符合算法所使用的假设时，现有的技术往往表现出良好的性能，但通常情况下，去噪算法不能完全去噪。可以通过增加x射线剂量值(超过医学上允许的最大剂量)来改善x射线图像质量，但这一过程可能对患者的健康是致命的，因为较高的x射线能量可能会由于较高剂量值的影响而杀死细胞。本研究将泊松主成分分析(Poisson Principal Component Analysis, PCA)与非局部均值去噪算法相结合，建立了一种混合模型来降低图像中的噪声。这种x射线噪声去除和对比度增强的混合模型提高了x射线图像的质量，因此可以用于医学诊断。用标准数据对混合模型的性能进行了观察，并与标准泊松算法进行了比较。
{"title":"Poisson Noise Reduction with Nonlocal-PCA Hybrid Model in Medical X-ray Images","authors":"Daniel Kipele, K. Greyson","doi":"10.18178/joig.11.2.178-184","DOIUrl":"https://doi.org/10.18178/joig.11.2.178-184","url":null,"abstract":"The presence of Poisson noise in medical X-ray images leads to degradation of the image quality. The obscured information is required for accurate diagnosis. During X-ray image acquisition process, weak light results into limited number of available photons, which leads into the Poisson noise commonly known as X-ray noise. Currently, the available X-ray noise removal methods have not yet obtained satisfying total denoising results to remove noise from the medical X-ray images. The available techniques tend to show good performance when the image model corresponds to the algorithm’s assumptions used but in general, the denoising algorithms fail to do complete denoise. X-ray image quality could be improved by increasing the X-ray dose value (beyond the maximum medically permissible dose) but the process could be lethal to patients’ health since higher X-ray energy may kill cells due to the effects of higher dose values. In this study, the hybrid model that combines the Poisson Principal Component Analysis (Poisson PCA) with the nonlocal (NL) means denoising algorithm is developed to reduce noise in images. This hybrid model for X-ray noise removal and the contrast enhancement improves the quality of X-ray images and can, thus, be used for medical diagnosis. The performance of the proposed hybrid model was observed by using the standard data and was compared with the standard Poisson algorithms.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76662502","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 : 2023-06-01DOI: 10.18178/joig.11.2.146-152
Libao Yang, S. Zenian, R. Zakaria
Fuzzy image enhancement is an important method in the process of image processing. Fuzzy image enhancement includes steps: gray-level fuzzification, modifying membership using intensifier (INT) operator, and obtaining new gray-levels by defuzzification. This paper proposed an adjustable INT operator with parameter k. Firstly, the image’s pixels are divided into two regions by the OTSU method (low and high region), and calculate the pixels’ membership by fuzzification in each region. Then, the INT operator reduce pixels’ membership in the low region and enlarge pixels’ membership in the high region. The parameter k is determined base on the pixel’s location information (neighborhood information), and plays an adjusting role when the INT operator is working. And finally, the result image is obtained by the defuzzification process. In the experiment results, the fuzzy image enhancement with the adjustable intensifier operator achieves a better performance.
{"title":"Fuzzy Image Enhancement Based on an Adjustable Intensifier OperatorFuzzy Image Enhancement Based on an Adjustable Intensifier Operator","authors":"Libao Yang, S. Zenian, R. Zakaria","doi":"10.18178/joig.11.2.146-152","DOIUrl":"https://doi.org/10.18178/joig.11.2.146-152","url":null,"abstract":"Fuzzy image enhancement is an important method in the process of image processing. Fuzzy image enhancement includes steps: gray-level fuzzification, modifying membership using intensifier (INT) operator, and obtaining new gray-levels by defuzzification. This paper proposed an adjustable INT operator with parameter k. Firstly, the image’s pixels are divided into two regions by the OTSU method (low and high region), and calculate the pixels’ membership by fuzzification in each region. Then, the INT operator reduce pixels’ membership in the low region and enlarge pixels’ membership in the high region. The parameter k is determined base on the pixel’s location information (neighborhood information), and plays an adjusting role when the INT operator is working. And finally, the result image is obtained by the defuzzification process. In the experiment results, the fuzzy image enhancement with the adjustable intensifier operator achieves a better performance.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85359081","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 : 2023-06-01DOI: 10.18178/joig.11.2.195-203
K. Saminathan, B. Sowmiya, Devi M Chithra
With increase in population, improving the quality and quantity of food is essential. Paddy is a vital food crop serving numerous people in various continents of the world. The yield of paddy is affected by numerous factors. Early diagnosis of disease is needed to prevent the plants from successive stage of disease. Manual diagnosis by naked eye is the traditional method widely adopted by farmers to identify leaf diseases. However, when the task involves manual disease diagnosis, problems like the hiring of domain experts, time consumption, and inaccurate results will arise. Inconsistent results may lead to improper treatment of plants. To overcome this problem, automatic disease diagnosis is proposed by researchers. This will help the farmers to accurately diagnose the disease swiftly without the need for expert. This manuscript develops model to classify four types of paddy leaf diseases bacterial blight, blast, tungro and brown spot. To begin with, the image is preprocessed by resizing and conversion to RGB Red, Green and Blue (RGB) and Hue, Saturation and Value (HSV) color space. Segmentation is done. Global features namely: hu moments, Haralick and color histogram are extracted and concatenated. Data is split in to training part and testing part in 70:30 ratios. Images are trained using multiple classifiers like Logistic Regression, Random Forest Classifier, Decision Tree Classifier, K-Nearest Neighbor (KNN) Classifier, Linear Discriminant Analysis (LDA),Support Vector Machine (SVM) and Gaussian Naive Bayes. This study reports Random Forest classifier as the best classifier. The Accuracy of the proposed model gained 92.84% after validation and 97.62% after testing using paddy disordered samples. 10 fold cross validation is performed. Performance of classification algorithms is measured using confusion matrix with precision, recall, F1- score and support as parameters.
{"title":"Multiclass Classification of Paddy Leaf Diseases Using Random Forest Classifier","authors":"K. Saminathan, B. Sowmiya, Devi M Chithra","doi":"10.18178/joig.11.2.195-203","DOIUrl":"https://doi.org/10.18178/joig.11.2.195-203","url":null,"abstract":"With increase in population, improving the quality and quantity of food is essential. Paddy is a vital food crop serving numerous people in various continents of the world. The yield of paddy is affected by numerous factors. Early diagnosis of disease is needed to prevent the plants from successive stage of disease. Manual diagnosis by naked eye is the traditional method widely adopted by farmers to identify leaf diseases. However, when the task involves manual disease diagnosis, problems like the hiring of domain experts, time consumption, and inaccurate results will arise. Inconsistent results may lead to improper treatment of plants. To overcome this problem, automatic disease diagnosis is proposed by researchers. This will help the farmers to accurately diagnose the disease swiftly without the need for expert. This manuscript develops model to classify four types of paddy leaf diseases bacterial blight, blast, tungro and brown spot. To begin with, the image is preprocessed by resizing and conversion to RGB Red, Green and Blue (RGB) and Hue, Saturation and Value (HSV) color space. Segmentation is done. Global features namely: hu moments, Haralick and color histogram are extracted and concatenated. Data is split in to training part and testing part in 70:30 ratios. Images are trained using multiple classifiers like Logistic Regression, Random Forest Classifier, Decision Tree Classifier, K-Nearest Neighbor (KNN) Classifier, Linear Discriminant Analysis (LDA),Support Vector Machine (SVM) and Gaussian Naive Bayes. This study reports Random Forest classifier as the best classifier. The Accuracy of the proposed model gained 92.84% after validation and 97.62% after testing using paddy disordered samples. 10 fold cross validation is performed. Performance of classification algorithms is measured using confusion matrix with precision, recall, F1- score and support as parameters.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84507094","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 : 2023-06-01DOI: 10.18178/joig.11.2.127-131
A. Saif, Z. R. Mahayuddin
Google Tilt Brush is expensive for virtual drawing which needs further improvement on the functionalities of mechanisms rather than implementation expects addressed in this research. Several issues are addressed by this research in this context, i.e., noise removal from sensor data, double integration-based drift issues and cost. Recently, available smart phones do not have the ability to perform drawing within artificial settings handling cardboard and daydream of google without purchasing Oculus Rift and HTC Vive (Virtual Reality Headset) because of expensiveness for large number of users. In addition, various extrinsic hardwares, i.e., satellite localization hardware and ultrasonic localization applications are not used for drawing in virtual reality. Proposed methodology implemented extended Kalman filter and Butterworth filter to perform positioning using six degree of freedom using Microelectromechanical Applications (MEMS) software data. A stereo visual method using Simultaneous Localization and Mapping (SLAM) is used to estimate the measurement for positioning implicating mobile phone (i.e., android platform) for the hardware system to estimate drift. This research implemented Google Virtual Reality application settings Kit with Unity3D engine. Experimentation validation states that proposed method can perform painting using virtual reality hardware integrated with controller software implicating smartphone mobile without using extrinsic controller device, i.e., Oculus Rift and HTC Vive with satisfactory accuracy.
{"title":"Stereo Vision Based Localization of Handheld Controller in Virtual Reality for 3D Painting Using Inertial System","authors":"A. Saif, Z. R. Mahayuddin","doi":"10.18178/joig.11.2.127-131","DOIUrl":"https://doi.org/10.18178/joig.11.2.127-131","url":null,"abstract":"Google Tilt Brush is expensive for virtual drawing which needs further improvement on the functionalities of mechanisms rather than implementation expects addressed in this research. Several issues are addressed by this research in this context, i.e., noise removal from sensor data, double integration-based drift issues and cost. Recently, available smart phones do not have the ability to perform drawing within artificial settings handling cardboard and daydream of google without purchasing Oculus Rift and HTC Vive (Virtual Reality Headset) because of expensiveness for large number of users. In addition, various extrinsic hardwares, i.e., satellite localization hardware and ultrasonic localization applications are not used for drawing in virtual reality. Proposed methodology implemented extended Kalman filter and Butterworth filter to perform positioning using six degree of freedom using Microelectromechanical Applications (MEMS) software data. A stereo visual method using Simultaneous Localization and Mapping (SLAM) is used to estimate the measurement for positioning implicating mobile phone (i.e., android platform) for the hardware system to estimate drift. This research implemented Google Virtual Reality application settings Kit with Unity3D engine. Experimentation validation states that proposed method can perform painting using virtual reality hardware integrated with controller software implicating smartphone mobile without using extrinsic controller device, i.e., Oculus Rift and HTC Vive with satisfactory accuracy.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81339470","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 : 2023-06-01DOI: 10.18178/joig.11.2.115-120
R. Radtke, Alexander Jesser
At the end of December 2019, a person in the Chinse city Wuhan was probably infected for the first time with the novel SARS-CoV-2 virus. In order to be able to react as quickly as possible after infection rapid diagnostic measures are of the utmost importance so that medical treatment can be taken at an early stage. An imageprocessing algorithm is presented using chest X-rays to determine whether a lung infection has a viral or a bacterial cause. In comparison to other more complicated evaluation methods, focus was put on using a simple algorithm by using the Canny algorithm for edge detection of infected areas of the lung tissue instead of complex deep learning processes. Main advantage here is that the method is portable to many different computer systems with little effort and does not need huge computing power. This should contribute to a faster diagnosis of SARS-CoV-2 virus-infection, especially in medically underdeveloped areas.
{"title":"Rapid Analysis of Thorax Images for the Detection of Viral Infections","authors":"R. Radtke, Alexander Jesser","doi":"10.18178/joig.11.2.115-120","DOIUrl":"https://doi.org/10.18178/joig.11.2.115-120","url":null,"abstract":"At the end of December 2019, a person in the Chinse city Wuhan was probably infected for the first time with the novel SARS-CoV-2 virus. In order to be able to react as quickly as possible after infection rapid diagnostic measures are of the utmost importance so that medical treatment can be taken at an early stage. An imageprocessing algorithm is presented using chest X-rays to determine whether a lung infection has a viral or a bacterial cause. In comparison to other more complicated evaluation methods, focus was put on using a simple algorithm by using the Canny algorithm for edge detection of infected areas of the lung tissue instead of complex deep learning processes. Main advantage here is that the method is portable to many different computer systems with little effort and does not need huge computing power. This should contribute to a faster diagnosis of SARS-CoV-2 virus-infection, especially in medically underdeveloped areas.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79292263","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}
Nowadays, humans can communicate easily with others by recognizing speech and text characters, particularly facial expressions. In human communication, it is critical to comprehend their emotion or implicit expression. Indeed, facial expression recognition is vital for analyzing the emotions of conversation partners, which can contribute to a series of matters, including mental health consulting. This technique enables psychiatrists to select appropriate questions based on their patients’ current emotional state. The purpose of this study was to develop a deep learningbased model for detecting and recognizing emotions on human faces. We divided the experiment into two parts: Faster R-CNN and mini-Xception architecture. We concentrated on four distinct emotional states: angry, sad, happy, and neutral. Both models implemented using the Faster R-CNN and the mini-Xception architectures were compared during the evaluation process. The findings indicate that the mini-Xception architecture model produced a better result than the Faster R-CNN. This study will be expanded in the future to include the detection of complex emotions such as sadness.
{"title":"Deep Learning-Based Emotion Recognition through Facial Expressions","authors":"Sarunya Kanjanawattana, Piyapong Kittichaiwatthana, Komsan Srivisut, Panchalee Praneetpholkrang","doi":"10.18178/joig.11.2.140-145","DOIUrl":"https://doi.org/10.18178/joig.11.2.140-145","url":null,"abstract":"Nowadays, humans can communicate easily with others by recognizing speech and text characters, particularly facial expressions. In human communication, it is critical to comprehend their emotion or implicit expression. Indeed, facial expression recognition is vital for analyzing the emotions of conversation partners, which can contribute to a series of matters, including mental health consulting. This technique enables psychiatrists to select appropriate questions based on their patients’ current emotional state. The purpose of this study was to develop a deep learningbased model for detecting and recognizing emotions on human faces. We divided the experiment into two parts: Faster R-CNN and mini-Xception architecture. We concentrated on four distinct emotional states: angry, sad, happy, and neutral. Both models implemented using the Faster R-CNN and the mini-Xception architectures were compared during the evaluation process. The findings indicate that the mini-Xception architecture model produced a better result than the Faster R-CNN. This study will be expanded in the future to include the detection of complex emotions such as sadness.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89893365","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 : 2023-06-01DOI: 10.18178/joig.11.2.132-139
Takato Sakai, M. Seo, N. Matsushiro, Yen-Wei Chen
The Yanagihara method is used to evaluate facial nerve palsy based on visual examinations by physicians. Examples of scored images are important for educational purposes and as references, however, due to patient privacy concern, actual facial images of real patients cannot be used for educational purposes. In this paper, we propose a solution to this problem by generating facial images of a virtual patient with facial nerve palsy, that can be shared and utilized by physicians. To reproduce the patient facial expression in a public face image, we propose a method to generate a swapped face image using the improved Cycle Generative Adversarial Networks (Cycle GAN) with a skiplayer excitation module and a self-supervised discriminator. Experimental results demonstrate that the proposed model can generate more coherent swapped faces that are similar to the public face identity and patient facial expressions. The proposed method also improves the quality of generated swapped face images while keeping them identical to the source (genuine) face image.
{"title":"Simulation of Facial Palsy Using Cycle GAN with Skip-Layer Excitation Module and Self-Supervised Discriminator","authors":"Takato Sakai, M. Seo, N. Matsushiro, Yen-Wei Chen","doi":"10.18178/joig.11.2.132-139","DOIUrl":"https://doi.org/10.18178/joig.11.2.132-139","url":null,"abstract":"The Yanagihara method is used to evaluate facial nerve palsy based on visual examinations by physicians. Examples of scored images are important for educational purposes and as references, however, due to patient privacy concern, actual facial images of real patients cannot be used for educational purposes. In this paper, we propose a solution to this problem by generating facial images of a virtual patient with facial nerve palsy, that can be shared and utilized by physicians. To reproduce the patient facial expression in a public face image, we propose a method to generate a swapped face image using the improved Cycle Generative Adversarial Networks (Cycle GAN) with a skiplayer excitation module and a self-supervised discriminator. Experimental results demonstrate that the proposed model can generate more coherent swapped faces that are similar to the public face identity and patient facial expressions. The proposed method also improves the quality of generated swapped face images while keeping them identical to the source (genuine) face image.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85620993","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 : 2023-03-01DOI: 10.18178/joig.11.1.15-20
Sachin Bahade, Michael Edwards, Xianghua Xie
Nuclei detection in histopathology images of cancerous tissue stained with conventional hematoxylin and eosin stain is a challenging task due to the complexity and diversity of cell data. Deep learning techniques have produced encouraging results in the field of nuclei detection, where the main emphasis is on classification and regressionbased methods. Recent research has demonstrated that regression-based techniques outperform classification. In this paper, we propose a classification model based on graph convolutions to classify nuclei, and similar models to detect nuclei using cascaded architecture. With nearly 29,000 annotated nuclei in a large dataset of cancer histology images, we evaluated the Convolutional Neural Network (CNN) and Graph Convolutional Networks (GCN) based approaches. Our findings demonstrate that graph convolutions perform better with a cascaded GCN architecture and are more stable than centre-of-pixel approach. We have compared our twofold evaluation quantitative results with CNN-based models such as Spatial Constrained Convolutional Neural Network (SC-CNN) and Centre-of-Pixel Convolutional Neural Network (CP-CNN). We used two different loss functions, binary cross-entropy and focal loss function, and also investigated the behaviour of CP-CNN and GCN models to observe the effectiveness of CNN and GCN operators. The compared quantitative F1 score of cascaded-GCN shows an improvement of 6% compared to state-of-the-art methods.
{"title":"Cascaded Graph Convolution Approach for Nuclei Detection in Histopathology Images","authors":"Sachin Bahade, Michael Edwards, Xianghua Xie","doi":"10.18178/joig.11.1.15-20","DOIUrl":"https://doi.org/10.18178/joig.11.1.15-20","url":null,"abstract":"Nuclei detection in histopathology images of cancerous tissue stained with conventional hematoxylin and eosin stain is a challenging task due to the complexity and diversity of cell data. Deep learning techniques have produced encouraging results in the field of nuclei detection, where the main emphasis is on classification and regressionbased methods. Recent research has demonstrated that regression-based techniques outperform classification. In this paper, we propose a classification model based on graph convolutions to classify nuclei, and similar models to detect nuclei using cascaded architecture. With nearly 29,000 annotated nuclei in a large dataset of cancer histology images, we evaluated the Convolutional Neural Network (CNN) and Graph Convolutional Networks (GCN) based approaches. Our findings demonstrate that graph convolutions perform better with a cascaded GCN architecture and are more stable than centre-of-pixel approach. We have compared our twofold evaluation quantitative results with CNN-based models such as Spatial Constrained Convolutional Neural Network (SC-CNN) and Centre-of-Pixel Convolutional Neural Network (CP-CNN). We used two different loss functions, binary cross-entropy and focal loss function, and also investigated the behaviour of CP-CNN and GCN models to observe the effectiveness of CNN and GCN operators. The compared quantitative F1 score of cascaded-GCN shows an improvement of 6% compared to state-of-the-art methods.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87525380","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}
Different breast cancer detection systems have been developed to help clinicians analyze screening mammograms. Breast cancer has been increasing gradually so scientists work to develop new methods to reduce the risks of this life-threatening disease. Convolutional Neural Networks (CNNs) have shown much promise In the field of medical imaging because of recent developments in deep learning. However, CNN’s based methods have been restricted due to the small size of the few public breast cancer datasets. This research has developed a new framework and introduced it to detect breast cancer. This framework utilizes Convolutional Neural Networks (CNNs) and image processing to achieve its goal because CNNs have been an important success in image recognition, reaching human performance. An efficient and fast CNN pre-trained object detector called RetinaNet has been used in this research. RetinaNet is an uncomplicated one-stage object detector. A two-stage transfer learning has been used with the selected detector to improve the performance. RetinaNet model is initially trained with a general-purpose dataset called COCO dataset. The transfer learning is then used to apply the RetinaNet model to another dataset of mammograms called the CBIS-DDSM dataset. Finally, the second transfer learning is used to test the RetinaNet model onto a small dataset of mammograms called the INbreast dataset. The results of the proposed two-stage transfer learning (RetinaNet → CBIS-DDSM → INbreast) are better than the other state-of-the-art methods on the public INbreast dataset. Furthermore, the True Positive Rate (TPR) is 0.99 ± 0.02 at 1.67 False Positives per Image (FPPI), which is better than the one-stage transfer learning with a TPR of 0.94 ± 0.02 at 1.67 FPPI.
已经开发了不同的乳腺癌检测系统来帮助临床医生分析筛查性乳房x光照片。乳腺癌一直在逐渐增加，因此科学家们致力于开发新的方法来降低这种危及生命的疾病的风险。由于深度学习的最新发展，卷积神经网络(cnn)在医学成像领域显示出很大的前景。然而，由于少数公开的乳腺癌数据集规模较小，CNN的基于方法受到了限制。这项研究开发了一种新的框架，并将其用于检测乳腺癌。该框架利用卷积神经网络(cnn)和图像处理来实现其目标，因为cnn在图像识别方面取得了重要的成功，达到了人类的表现。在这项研究中使用了一种高效快速的CNN预训练对象检测器，称为RetinaNet。retanet是一个简单的单级目标探测器。采用两阶段迁移学习方法对所选择的检测器进行学习，以提高性能。retanet模型最初使用一个称为COCO数据集的通用数据集进行训练。然后使用迁移学习将RetinaNet模型应用于另一个称为CBIS-DDSM数据集的乳房x光片数据集。最后，第二次迁移学习用于在一个称为INbreast数据集的乳房x光照片小数据集上测试RetinaNet模型。所提出的两阶段迁移学习(RetinaNet→CBIS-DDSM→INbreast)在公共INbreast数据集上的结果优于其他最先进的方法。在1.67个False Positives per Image (FPPI)下，该方法的True Positive Rate (TPR)为0.99±0.02，优于单阶段迁移学习(1.67个FPPI)下的TPR(0.94±0.02)。
{"title":"Breast Cancer Detection Using Image Processing and Machine Learning","authors":"Z. Q. Habeeb, B. Vuksanovic, Imad Al-Zaydi","doi":"10.18178/joig.11.1.1-8","DOIUrl":"https://doi.org/10.18178/joig.11.1.1-8","url":null,"abstract":"Different breast cancer detection systems have been developed to help clinicians analyze screening mammograms. Breast cancer has been increasing gradually so scientists work to develop new methods to reduce the risks of this life-threatening disease. Convolutional Neural Networks (CNNs) have shown much promise In the field of medical imaging because of recent developments in deep learning. However, CNN’s based methods have been restricted due to the small size of the few public breast cancer datasets. This research has developed a new framework and introduced it to detect breast cancer. This framework utilizes Convolutional Neural Networks (CNNs) and image processing to achieve its goal because CNNs have been an important success in image recognition, reaching human performance. An efficient and fast CNN pre-trained object detector called RetinaNet has been used in this research. RetinaNet is an uncomplicated one-stage object detector. A two-stage transfer learning has been used with the selected detector to improve the performance. RetinaNet model is initially trained with a general-purpose dataset called COCO dataset. The transfer learning is then used to apply the RetinaNet model to another dataset of mammograms called the CBIS-DDSM dataset. Finally, the second transfer learning is used to test the RetinaNet model onto a small dataset of mammograms called the INbreast dataset. The results of the proposed two-stage transfer learning (RetinaNet → CBIS-DDSM → INbreast) are better than the other state-of-the-art methods on the public INbreast dataset. Furthermore, the True Positive Rate (TPR) is 0.99 ± 0.02 at 1.67 False Positives per Image (FPPI), which is better than the one-stage transfer learning with a TPR of 0.94 ± 0.02 at 1.67 FPPI.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83824098","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 : 2023-03-01DOI: 10.18178/joig.11.1.61-71
L. Reyes-Ruiz, duardo Fragoso-Navarro, F. Garcia-Ugalde, O. Juarez-Sandoval, M. Cedillo-Hernández, M. Nakano-Miyatake
Nowadays, advances in information and communication technologies along with easy access to electronic devices such as smartphones have achieved an agile and efficient storing, edition, and distribution of digital multimedia files. However, lack of regulation has led to several problems associated with intellectual property authentication and copyright protection. Furthermore, the problem becomes complex in a scenario of illegal printed exploitation, which involves printing and scanning processes. To solve these problems, several digital watermarking in combination with cryptographic algorithms has been proposed. In this paper, a strategy of robust watermarking is defined consisting of the administration and detection of unauthorized use of digitized cinematographic images from Mexican cultural heritage. The proposed strategy is based on the combination of two types of digital watermarking, one of visible-camouflaged type based on spatial domain and another of invisible type based on frequency domain, together with a particle swarm optimization. The experimental results show the high performance of the proposed algorithm faced to printing-scanning processes or digital-analogue attack, and common image geometric and image processing attacks such as JPEG compression. Additionally, the imperceptibility of the watermark is evaluated by PSNR and compared with other previously proposed algorithms.
{"title":"Robust Dual Digital Watermark Applied to Antique Digitized Cinema Images: Resistant to Print-Scan Attack","authors":"L. Reyes-Ruiz, duardo Fragoso-Navarro, F. Garcia-Ugalde, O. Juarez-Sandoval, M. Cedillo-Hernández, M. Nakano-Miyatake","doi":"10.18178/joig.11.1.61-71","DOIUrl":"https://doi.org/10.18178/joig.11.1.61-71","url":null,"abstract":"Nowadays, advances in information and communication technologies along with easy access to electronic devices such as smartphones have achieved an agile and efficient storing, edition, and distribution of digital multimedia files. However, lack of regulation has led to several problems associated with intellectual property authentication and copyright protection. Furthermore, the problem becomes complex in a scenario of illegal printed exploitation, which involves printing and scanning processes. To solve these problems, several digital watermarking in combination with cryptographic algorithms has been proposed. In this paper, a strategy of robust watermarking is defined consisting of the administration and detection of unauthorized use of digitized cinematographic images from Mexican cultural heritage. The proposed strategy is based on the combination of two types of digital watermarking, one of visible-camouflaged type based on spatial domain and another of invisible type based on frequency domain, together with a particle swarm optimization. The experimental results show the high performance of the proposed algorithm faced to printing-scanning processes or digital-analogue attack, and common image geometric and image processing attacks such as JPEG compression. Additionally, the imperceptibility of the watermark is evaluated by PSNR and compared with other previously proposed algorithms.","PeriodicalId":36336,"journal":{"name":"中国图象图形学报","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81786159","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：