KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 10 of 12

Some of the metrics are blocked by your
consent settings
Ensemble deep learning for the detection of COVID-19 in unbalanced chest X-ray dataset
(2021-11-01)
Win, Khin Yadanar
;
Maneerat, Noppadol
;
Sreng, Syna
;
Hamamoto, Kazuhiko
The ongoing COVID-19 pandemic has caused devastating effects on humanity worldwide. With practical advantages and wide accessibility, chest X-rays (CXRs) play vital roles in the diagnosis of COVID-19 and the evaluation of the extent of lung damages incurred by the virus. This study aimed to leverage deep-learning-based methods toward the automated classification of COVID-19 from normal and viral pneumonia on CXRs, and the identification of indicative regions of COVID-19 biomarkers. Initially, we preprocessed and segmented the lung regions usingDeepLabV3+ method, and subsequently cropped the lung regions. The cropped lung regions were used as inputs to several deep convolutional neural networks (CNNs) for the prediction of COVID-19. The dataset was highly unbalanced; the vast majority were normal images, with a small number of COVID-19 and pneumonia images. To remedy the unbalanced distribution and to avoid biased classification results, we applied five different approaches: (i) balancing the class using weighted loss; (ii) image augmentation to add more images to minority cases; (iii) the undersampling of majority classes; (iv) the oversampling of minority classes; and (v) a hybrid resampling approach of oversampling and undersampling. The best-performing methods from each approach were combined as the ensemble classifier using two voting strategies. Finally, we used the saliency map of CNNs to identify the indicative regions of COVID-19 biomarkers which are deemed useful for interpretability. The algorithms were evaluated using the largest publicly available COVID-19 dataset. An ensemble of the top five CNNs with image augmentation achieved the highest accuracy of 99.23% and area under curve (AUC) of 99.97%, surpassing the results of previous studies.
Some of the metrics are blocked by your
consent settings
Hybrid learning of hand-crafted and deep-activated features using particle swarm optimization and optimized support vector machine for tuberculosis screening
(2020-09-01)
Win, Khin Yadanar
;
Maneerat, Noppadol
;
Hamamoto, Kazuhiko
;
Sreng, Syna
Tuberculosis (TB) is a leading infectious killer, especially for people with Human Immunodeficiency Virus (HIV) and Acquired Immunodeficiency Syndrome (AIDS). Early diagnosis of TB is crucial for disease treatment and control. Radiology is a fundamental diagnostic tool used to screen or triage TB. Automated chest x-rays analysis can facilitate and expedite TB screening with fast and accurate reports of radiological findings and can rapidly screen large populations and alleviate a shortage of skilled experts in remote areas. We describe a hybrid feature-learning algorithm for automatic screening of TB in chest x-rays: it first segmented the lung regions using the DeepLabv3+ model. Then, six sets of hand-crafted features from statistical textures, local binary pattern, GIST, histogram of oriented gradients (HOG), pyramid histogram of oriented gradients and bags of visual words (BoVW), and nine sets of deep-activated features from AlexNet, GoogLeNet, InceptionV3, XceptionNet, ResNet-50, SqueezeNet, ShuffleNet, MobileNet, and DenseNet, were extracted. The dominant features of each feature set were selected using particle swarm optimization, and then separately input to an optimized support vector machine classifier to label 'normal' and 'TB' x-rays. GIST, HOG, BoVW from hand-crafted features, and MobileNet and DenseNet from deep-activated features performed better than the others. Finally, we combined these five best-performing feature sets to build a hybrid-learning algorithm. Using the Montgomery County (MC) and Shenzen datasets, we found that the hybrid features of GIST, HOG, BoVW, MobileNet and DenseNet, performed best, achieving an accuracy of 92.5% for the MC dataset and 95.5% for the Shenzen dataset.
Some of the metrics are blocked by your
consent settings
Deep learning for optic disc segmentation and glaucoma diagnosis on retinal images
(2020-07-01)
Sreng, Syna
;
Maneerat, Noppadol
;
Hamamoto, Kazuhiko
;
Win, Khin Yadanar
Glaucoma is a major global cause of blindness. As the symptoms of glaucoma appear, when the disease reaches an advanced stage, proper screening of glaucoma in the early stages is challenging. Therefore, regular glaucoma screening is essential and recommended. However, eye screening is currently subjective, time-consuming and labor-intensive and there are insufficient eye specialists available. We present an automatic two-stage glaucoma screening system to reduce the workload of ophthalmologists. The system first segmented the optic disc region using a DeepLabv3+ architecture but substituted the encoder module with multiple deep convolutional neural networks. For the classification stage, we used pretrained deep convolutional neural networks for three proposals (1) transfer learning and (2) learning the feature descriptors using support vector machine and (3) building ensemble of methods in (1) and (2). We evaluated our methods on five available datasets containing 2787 retinal images and found that the best option for optic disc segmentation is a combination of DeepLabv3+ and MobileNet. For glaucoma classification, an ensemble of methods performed better than the conventional methods for RIM-ONE, ORIGA, DRISHTI-GS1 and ACRIMA datasets with the accuracy of 97.37%, 90.00%, 86.84% and 99.53% and Area Under Curve (AUC) of 100%, 92.06%, 91.67% and 99.98%, respectively, and performed comparably with CUHKMED, the top team in REFUGE challenge, using REFUGE dataset with an accuracy of 95.59% and AUC of 95.10%.
Some of the metrics are blocked by your
consent settings
Cotton wool spots detection in diabetic retinopathy based on adaptive thresholding and ant colony optimization coupling support vector machine
(2019-06-01)
Sreng, Syna
;
Maneerat, Noppadol
;
Hamamoto, Kazuhiko
;
Panjaphongse, Ronakorn
Diabetic retinopathy is the major issue of diabetes-induced blindness worldwide but is curable if detected in time. Cotton wool spots (CWSs) are the critical lesions of diabetic retinopathy, which indicate not only advanced nonproliferative but also preproliferative diabetic retinopathy. It is crucial to detect CWSs for grading the severity of diabetic retinopathy. By grading the severity of diabetic retinopathy accurately, the eye specialist can make an effective treatment plan to protect the patient's vision against blindness. CWSs detection remains challenging because of their uneven appearance, in which some CWSs are not clearly visible and some resemble hard exudates. This paper proposed an automatic CWS detection method based on adaptive thresholding and ant colony optimization (ACO) coupled with support vector machine (SVM). One-hundred and sixty-two features from five feature sets, namely morphologies, first-order statistics, gray-level co-occurrence matrix, gray-level run length matrix, and lacunarity, are extracted, and then four feature selection methods,namely genetic algorithm, particle swarm optimization, stepwise method, and ACO, are coupled with SVM classifiers. The evaluation results of the proposed methods on local, standard diabetic retinopathy database calibration level 1, and high-resolution fundus image database datasets containing 319 images indicate that ACO coupling cubic SVM performs better than the other pairs with sensitivity 90.16%, specificity 97.92%, accuracy 96.96%, and area under receiver operating characteristic curve 97.19%. © 2019 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.
Some of the metrics are blocked by your
consent settings
Classification of Cotton Wool Spots Using Principal Components Analysis and Support Vector Machine
(2019-01-10)
Sreng, Syna
;
Maneerat, Noppadol
;
Win, Khin Yadanar
;
Hamamoto, Kazuhiko
;
Panjaphongse, Ronakorn
Diabetic retinopathy is a complication of the eye damage and can lead to being blindness if it is late for treatment. Microaneurysms, exudates, hemorrhages and cotton wool spots are the lesions associated with diabetic retinopathy. Numerous studies have been done on the detection of microaneurysms, and hemorrhages, as well as exudates whereas only a few research works for detection of cotton wool spots, mainly because of the fact that its appearances are difficult to filter out from the background and not clearly visible. In this paper, an algorithm is proposed to detect cotton wool spots based on integrating principal components analysis and support vector machine. First, preprocessing is performed to enhance the retinal images. Then adaptive thresholding method is used to roughly extract the cotton wool spot from the background. Support vector machine and principal components analysis are further applied respectively to select the important features from morphologies, first-order statistics, gray level occurrence matrix and lacunarity. The proposed method was evaluated with local and DIARETDB1 datasets containing 289 images. Given a success rate of accuracy 90.47 %, sensitivity 85.29%, and specificity 90.12% with the average computational time 16.47 seconds per image on cotton wool spots detection, this system performed better by comparing to the previous research works.
Some of the metrics are blocked by your
consent settings
Suitable Supervised Machine Learning Techniques for Malignant Mesothelioma Diagnosis
(2019-01-10)
Win, Khin Yadanar
;
Maneerat, Noppadol
;
Choomchuay, Somsak
;
Sreng, Syna
;
Hamamoto, Kazuhiko
Malignant Mesothelioma (MM) is a rare, aggressive cancer that grows in the lining of the internal organs such as lung, abdomen or heart. Fousing on MM diagnosis, in this paper, we investigate multiple machine learning methods and compare for accurate MM diagnosis results. Seven machine learning algorithms namely (i) Linear Discriminant Analysis (LDA), (ii) Naïve Bayes, (iii) K Nearest Neighborhood (KNN), (iv) Support Vector Machine (SVM), (v) Decision Tree (DT), (vi) Logistic Regression (LogR) and (vii) Random forest (RF) algorithms are exploited. The experiments dataset containing 324 cases with 34 features and six performance measures are used to assess the accuracy of evaluated classifiers. The average accuracy of LDA, NB, KNN, SVM, DT, LogR and RF are 61.73%, 67.90%, 91.36%, 100%, 100%, 100% and 100%, respectively. In addition, the computational complexity of each method is also analyzed. Each algoritm is judged based on its classification accuracy and computational complexity. It is found that SVM, DT, LogR and RF outperform the others and even previous studies.
Some of the metrics are blocked by your
consent settings
Automated diabetic retinopathy screening system using hybrid simulated annealing and ensemble bagging classifier
(2018-07-22)
Sreng, Syna
;
Maneerat, Noppadol
;
Hamamoto, Kazuhiko
;
Panjaphongse, Ronakorn
Diabetic Retinopathy (DR) is the leading cause of blindness in working-age adults globally. Primary screening of DR is essential, and it is recommended that diabetes patients undergo this procedure at least once per year to prevent vision loss. However, in addition to the insufficient number of ophthalmologists available, the eye examination itself is labor-intensive and time-consuming. Thus, an automated DR screening method using retinal images is proposed in this paper to reduce the workload of ophthalmologists in the primary screening process and so that ophthalmologists may make effective treatment plans promptly to help prevent patient blindness. First, all possible candidate lesions of DR were segmented from the whole retinal image using a combination of morphological-top-hat and Kirsch edge-detection methods supplemented by pre- and post-processing steps. Then, eight feature extractors were utilized to extract a total of 208 features based on the pixel density of the binary image as well as texture, color, and intensity information for the detected regions. Finally, hybrid simulated annealing was applied to select the optimal feature set to be used as the input to the ensemble bagging classifier. The evaluation results of this proposed method, on a dataset containing 1200 retinal images, indicate that it performs better than previous methods, with an accuracy of 97.08%, a sensitivity of 90.90%, a specificity of 98.92%, a precision of 96.15%, an F-measure of 93.45% and the area under receiver operating characteristic curve at 98.34%.
Some of the metrics are blocked by your
consent settings
Primary screening of diabetic retinopathy based on integrating morphological operation and support vector machine
(2017-07-02)
Sreng, Syna
;
Maneerat, Noppadol
;
Isarakorn, Don
;
Hamamoto, Kazuhiko
;
Panjaphongse, Ronakorn
Diabetic retinopathy is one of the most frequent causes of blindness due to diabetes. Primary screening is essential due to prerequisite step toward the diagnosis of diabetic retinopathy in order to prevent vision loss or blindness. This paper presents the methods to discriminate between healthy images and diabetic retinopathy images on the retinal images. The proposed method involves three main steps. Initially, the image is preprocessed to remove small noises and enhance the contrast of the image. Secondly, Kirsch edge detection is utilized to detect the bright lesions. Subsequently, the red lesions are detected depending on top-hat morphological filtering methods. Then the bright and dark lesions are combined by using logical AND operator. In order to be left only pathological signs, the noises near the vicinity of the optic disc and blood vessels are further removed using blob analysis. Finally, morphological features are extracted and fed to the SVM classifier. The proposed method was evaluated with three datasets containing 229 images. It achieved the accuracy of 90%, sensitivity of 86.33% and specificity of 98.55% with the average computational time 8 seconds per image. The method is simple and fast, easy to implement and the result is promising.
Some of the metrics are blocked by your
consent settings
Automated microaneurysms detection in fundus images using image segmentation
(2017-04-19)
Sreng, Syna
;
Maneerat, Noppadol
;
Hamamoto, Kazuhiko
Diabetic retinopathy is one of the complicated diseases which occurs in diabetic patients when the affects damage the retina. The eyes vision can lead to be lost in case of late treatment. Microaneurysms are the earliest detectable abnormalities of diabetic retinopathy, so the automated detection of the lesions is essential and useful task. This paper proposed a simple method to detect microaneurysms based on its characteristics in fundus images using some techniques in image segmentation. First, we preprocessed to reduce image noise and improve the contrast. Then we segmented them using Canny edges detection and maximum entropy thresholding. The characteristics of microaneurysms which appear as small red dots and circular shape are the specific points to discriminate them from the other lesions as well as the anatomical structures of the fundus image by applying area and eccentricity methods. Finally, the morphological operation was applied to mark out these symptoms. The results were analysis by ophthalmologist in order to define system accuracy and preciseness. According to results of comparison, we found that the accuracy is 90 % and the average processing time is 9.53 seconds per image.
Some of the metrics are blocked by your
consent settings
Automatic hemorrhages detection based on fundus images
(2015-01-01)
Sreng, Syna
;
Maneerat, Noppadol
;
Isarakorn, Don
;
Hamamoto, Kazuhiko
;
Panjaphongse, Ronakorn
This paper proposes methods to detect hemorrhages which are known as a kind of lesions in diabetic retinopathy. To detect the symptom, eye fundus structures (blood vessels and fovea) as well as microaneuysms need to be discriminated to filter out only the hemorrhages. Five processing steps are proposed based analysis on fundus images. First, preprocessing step is processed to improve the quality of the image. Then all red features are filtered out. They include blood vessels, fovea, microaneurysms and hemorrhages. After that, morphology operation and compactness measurement are applied to eliminate the fovea, and blood vessels. Finally, hemorrhages can be classified by using area method to remove microaneurysms and some small noise. 579 fundus images from Bhumibol Adulyadej Hospital were tested. The results were analysis by ophthalmologist in order to define system accuracy and preciseness. According to results of comparison, we found that the accuracy is 90 % and the average of processing time is 6.23 seconds per image.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results