KMITL

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

Browse

Search Results

Now showing 1 - 10 of 22

Some of the metrics are blocked by your
consent settings
Enhanced deep learning model for prediction of diabetic mellitus on optical coherence tomography angiography images
(2026-01-01)
Visitsattapongse, Sarinporn
;
Rithcharung, Preeyarat
;
Santiprabhob, Jeerunda
;
Lertbannaphong, Ornsuda
;
Sermsripong, Wasawat
Background: Diabetes mellitus (DM) is a chronic metabolic disease characterized by dysregulated blood glucose. Prolonged DM can lead to diabetic retinopathy (DR), in which retinal capillaries are damaged by sustained hyperglycemia. Optical coherence tomography angiography (OCTA) is a non-invasive imaging modality for visualizing retinal microvasculature and can detect early changes in both DM patients with and without DR. However, it requires expert evaluation, making early detection costly and time-consuming. This study aimed to develop a high-performance deep learning framework that can classify OCTA images into three groups of DM, such as normal, good glycemic control, and poor glycemic control. Methods: OCTA datasets of horizontal B-scans and en face scans from 300 participants aged 8–18 years were analyzed, including normal controls, DM patients with good glycemic control, and DM patients with poor control (HbA1c ≥8%). For each participant, a 3 mm × 3 mm foveal-centered en face image of the deep capillary plexus (DCP) and a horizontal B-scan through the foveal center of the right eye were selected. Several convolutional and transformer-based models were evaluated, with ConvNeXt (a ConvNet for the 2020s) chosen as the baseline for its superior performance. To enhance generalization and convergence, progressive resizing and the Lookahead optimization strategy were applied, while class-wise augmentation was used to balance the training set without altering the test distribution. Results: The baseline ConvNeXt achieved F1 scores of 0.7877 (B-scans) and 0.7424 (en face). After doing enhancement using progressive resizing and Lookahead optimization, performance improved to 0.8319 and 0.8567 (Wilcoxon signed-rank tests, P<0.05). Conclusions: Our proposed method for DM classification from OCTA images provided promising results while ensuring resource efficiency and rapid evaluation. Clinically, accurate classification of DM status is valuable for assessing the risk of DR progression. Thus, it can be served as an assistive tool for clinical decision support in DR management.
Some of the metrics are blocked by your
consent settings
Analyzing explainability of YOLO-based breast cancer detection using heat map visualizations
(2025-07-01)
Ariyametkul, Awika
;
Paing, May Phu
Background: Breast cancer is the most frequently diagnosed and leading cause of cancer-related mortality among women worldwide. The danger of this disease is due to its asymptomatic nature in the early stages, thereby underscoring the importance of early detection. Mammography, a specialized X-ray imaging technique for breast examination, has been pivotal in facilitating early detection and reducing mortality rates. In recent years, artificial intelligence (AI) has gained substantial popularity across various fields, including medicine. Numerous studies have leveraged AI techniques, particularly convolutional neural networks (CNNs) and You Only Look Once (YOLO)-based models, for medical image detection and classification. However, the predictions of such AI models often lack transparency and explainability, resulting in low trustworthiness. This study aims to address this gap by investigating three state-of-the-art versions of the YOLO algorithm—YOLO version 9 (YOLOv9), YOLO version 10 (YOLOv10), and YOLO version 11 (YOLO11)—trained on breast cancer imaging datasets, specifically the INbreast and Mammographic Image Analysis Society (MIAS) databases. Additionally, to address the challenges posed by the lack of explainability and transparency, we integrate seven explainable artificial intelligence (XAI) methods: Grad-CAM, GradCAM++, Eigen-CAM, EigenGrad-CAM, XGrad-CAM, LayerCAM, and HiResCAM. Methods: This study utilized two publicly available breast cancer image databases: INbreast: toward a Full-field Digital Mammographic Database and the MIAS dataset. Preprocessing steps were applied to standardize all images in accordance with the input requirements of the YOLO architecture, as these datasets were used to train the three most recent versions of YOLO. The YOLO model demonstrating the highest performance—measured by mean average precision (mAP), precision, and recall—was selected for integration with seven different XAI methods. The performance of each XAI technique was evaluated both qualitatively through visual inspection and quantitatively using several metrics, including matching ground truth (mGT), Pearson correlation coefficient (PCC), precision, recall, and root mean square error (RMSE). These methodologies were employed to interpret and visualize the “black box” decision-making processes of the top-performing YOLO model. Results: Based on our experimental findings, YOLO11 outperformed YOLOv9 (mAP 0.868) and YOLOv10 (mAP 0.926), achieving the highest mAP of 0.935, with classification accuracies of 95% for benign and 80% for malignant cases. Among the evaluated XAI techniques, HiResCAM provided the most effective visual explanations, attaining the highest mGT score of 0.49, surpassing EigenGrad-CAM (0.45) and LayerCAM (0.42) in both visual and quantitative evaluations. Conclusions: The integration of YOLO11 with HiResCAM offers a robust solution that combines high detection accuracy with improved model interpretability. This approach not only enhances user trustworthiness by revealing decision-making patterns and limitations but also provide insights into the weaknesses of the model, enabling developers to refine and improve AI performance further.
Some of the metrics are blocked by your
consent settings
Comparing AI Decision-Making with Expert Biomarkers: A Case Study on Diabetic Retinopathy Classification
(2025-01-01)
Sharma, Aayushma
;
Paing, May Phu
Artificial intelligence (AI) has become prevalent in the healthcare sector due to its ability to interpret complex medical images that may not be apparent to humans. Traditional black-box models were used to classify the disease, providing no information as to why certain things were labeled as such and not others. This paper utilizes the use of eXplainable-AI (XAI), specifically, Layer-wise Relevance Propagation (LRP) which generates mapping between AI decision and biomarker used by the ophthalmologist whereby enhancing results interpretability and transparency in the disease diagnostic tasks. VGG-16 incorporated with batch normalization and label smoothing was used for the classification tasks whereas LRP was employed to perform the heat-map generation to see if the feature extracted and used by AI was consistent with the experts' biomarkers. Our proposed model obtained a classification accuracy of 77.33%, where 165 out of 266 images were aligned with the ophthalmologist's prediction. Furthermore, the significance of heatmap generation was supported by a one-sample Z-test which revealed that the alignment between AI predictions and expert biomarkers is significantly greater than random, with a 95% confidence interval.
Some of the metrics are blocked by your
consent settings
Advancing Breast Cancer Identification: Exploring Deep Learning Models for Improved Detection
(2025-01-01)
Samrankit, Thitiphon
;
Surakiat, Kanyanut
;
Tamang, Sudarshan
;
Sharma, Aayushma
;
Paing, May Phu
(1) Background: Breast cancer is the uncontrolled growth of abnormal cells within the breast tissue, and it can be either malignant or benign. Malignant tumors are cancerous and have the potential to invade and spread to other parts of the body while benign tumors are non-cancerous. Traditional breast cancer screening methods have limitations from artifacts and positioning errors. (2) Methods: This study employs various deep learning models—GELAN-C (Generalized Efficient Layer Aggregation Network – Compact version), YOLOv8 (You Only Look Once – version 8), RTMDet (Real-Time Models for Object Detection), DETR (DEtection TRansformer), YOLO-NAS (You Only Look Once – Neural Architecture Search), and Detectron2—to identify the most effective and adaptable model for breast cancer diagnosis. (3) Results: From our experiment, GELAN- C outperformed other models providing exceptional accuracy with the highest mean average precision (mAP) scores of both thresholds. Specifically, GELAN-C achieved a mAP@0.50 of 0.983, which reflects overall object detection accuracy. Additionally, it maintained a strong mAP@0.50-0.95 of 0.868 indicating its robustness and precision across varying levels of detection difficulty.
Some of the metrics are blocked by your
consent settings
Instance segmentation of cells and nuclei from multi-organ cross-protocol microscopic images
(2024-09-01)
Baral, Sushish
;
Paing, May Phu
Background: Light microscopy is a widely used technique in cell biology due to its satisfactory resolution for cellular structure analysis, prevalent availability of fluorescent probes for staining, and compatibility for the dynamic analysis of live cells. However, the segmentation of cells and nuclei from microscopic images is not a straightforward process because it has several challenges such as high variation in morphology and shape, the presence of noise and diverse contrast in backgrounds, clustering or overlapping nature of cells. Dealing with these challenges and facilitating more reliable analysis necessitates the implementation of computer-aided methods that leverage image processing techniques and deep learning algorithms. The major goal of this study is to propose a model, for instance segmentation of cells and nuclei, applying the most cutting-edge deep learning techniques. Methods: A fine-tuned You Only Look at Once version 9 extended (YOLOv9-E) model is initially applied as a prompt generator to generate bounding box prompts. Using the generated prompts, a pre-trained segment anything model (SAM) is subsequently applied through zero-short inferencing to produce raw segmentation masks. These segmentation masks are then refined using non-max suppression and simple image processing methods such as image addition and morphological processing. The proposed method is developed and evaluated using an open-sourced dataset called Expert Visual Cell Annotation (EVICAN), which is relatively large and contains 4,738 microscopy images extracted from cross organs using different protocols. Results: Based on the evaluation results on three different levels of EVICAN test sets, the proposed method demonstrates noticeable performances showing average mAP50 [mean average precision at intersection over union (IoU) =0.50] scores of 96.25, 95.05, and 94.18 for cell segmentation, and 68.04, 54.66, and 38.29 for nucleus segmentation on easy, medium, and difficult test sets, respectively. Conclusions: Our proposed method for instance segmentation of cells and nuclei provided favorable performance compared to the existing methods in literature, indicating its potential utility as an assistive tool for cell culture experts, facilitating prompt and reliable analysis.
Some of the metrics are blocked by your
consent settings
Explainable AI (XAI) for Breast Cancer Diagnosis
(2024-01-01)
Ariyametkul, Awika
;
Tamang, Sudarshan
;
Paing, May Phu
Breast cancer is the leading cause of mortality and incidence among women worldwide. Mammography, an essential imaging technique, plays a pivotal role in both screening and diagnostic processes by facilitating early detection, which helps improve survival rates. Despite its effectiveness, interpreting mammographic images presents considerable challenges, necessitating the expertise of highly trained radiologists. Artificial intelligence (AI) is a powerful tool for managing large amounts of data and is increasingly used across numerous sectors, including medical applications. This research focuses on applying Convolutional Neural Networks (CNNs) to classify breast cancer from mammograms. We explored six different CNN models including simple ConvNet, AlexNet, VGG-16, GoogLeNet, XceptionNet, and DenseNet201. Our results indicate that DenseNet201 is the most suitable model for this task, achieving 99% accuracy. However, a limitation of AI is the lack of transparency and explanation, often referred to as the 'black box' problem. This vulnerability can be addressed through explainable artificial intelligence (XAI), which elucidates the processes behind AI's decision-making. We employed three different XAI methodologies, including LIME, GradCAM, and GradCAM++, to visualize the model's decision-making process.
Some of the metrics are blocked by your
consent settings
Deep Learning for Segmentation of Brain Tumors
(2024-01-01)
Tamang, Sudarshan
;
Ariyametkul, Awika
;
Paing, May Phu
;
Bui, Toan H.
(1) Background: Meningioma, Glioma, and Pituitary tumors are some of the three most common brain tumors that have great adverse effects on normal functionalities of the brain. These tumors are hard to detect and take a significant amount of time manually. To overcome this problem, deep learning techniques help in automating tumor detection quickly. This helps in early detection of brain tumors and patients can receive treatment before the tumor gets worse. (2) Methods: An automated segmentation model was created using UNet as a base model along variation of ResNet architecture. Magnetic resonance imaging (MRI) scans with T1-weighted contrast-enhanced images having 128 × 128 pixels in dimensions, are categorized into 3 classes- Meningioma (MEN), Glioma (GLI), and Pituitary tumor (PIT). At last, the predicted tumor by the proposed model was compared with the ground truth label of the corresponding tumor class. Dice score and IoU were used as performance metrics of the model. (3) Results: In this study, the best performing model was ResNeXt50_32x4dUNet among the 7 models used. The mean test dice score from this model was 0.835 whereas the mean validation dice score was 0. 784. Similarly, the mean IoU of the test images was 73.0% and showed an acceptable performance in the segmentation of brain tumor.
Some of the metrics are blocked by your
consent settings
A Deep Learning Model for Bacterial Classification Using Big Transfer (BiT)
(2024-01-01)
Visitsattaponge, Sarinporn
;
Bunkum, Manao
;
Pintavirooj, Chuchart
;
Paing, May Phu
Identification and classification of bacterial genera and species are very important for medical prevention, diagnosis, and treatment. However, due to microbial diversity and high variability in appearance, the manual classification of bacteria is a challenging and time-consuming task. This paper aims to facilitate such a troublesome task using deep learning techniques. Through the utilization of a deep learning model, specifically a Big Transfer (BiT) combined with graph Laplacian-based data cleaning and weight initialization based-rectified linear unit (WIB-Relu) activation, we have developed an accurate bacteria classification model. We have tested our proposed method on a public dataset of microscopic bacteria images, called the Digital Images of Bacteria Species (DIBaS), and achieved promising results with an accuracy of 99.11%, precision of 99.31%, recall of 99.09%, and F1 score of 99.06%, respectively. Moreover, the proposed bacteria classification performed well regardless of the size of the training data. We investigated its generalizability not only on the original dataset but also on the few shots (5-shots, 2-shots, and 1-shot) and augmented datasets.
Some of the metrics are blocked by your
consent settings
Parking Time Violation Tracking Using YOLOv8 and Tracking Algorithms
(2023-07-01)
Sharma, Nabin
;
Baral, Sushish
;
Paing, May Phu
;
Chawuthai, Rathachai
The major problem in Thailand related to parking is time violation. Vehicles are not allowed to park for more than a specified amount of time. Implementation of closed-circuit television (CCTV) surveillance cameras along with human labor is the present remedy. However, this paper presents an approach that can introduce a low-cost time violation tracking system using CCTV, Deep Learning models, and object tracking algorithms. This approach is fairly new because of its appliance of the SOTA detection technique, object tracking approach, and time boundary implementations. YOLOv8, along with the DeepSORT/OC-SORT algorithm, is utilized for the detection and tracking that allows us to set a timer and track the time violation. Using the same apparatus along with Deep Learning models and algorithms has produced a better system with better performance. The performance of both tracking algorithms was well depicted in the results, obtaining MOTA scores of (1.0, 1.0, 0.96, 0.90) and (1, 0.76, 0.90, 0.83) in four different surveillance data for DeepSORT and OC-SORT, respectively.
Some of the metrics are blocked by your
consent settings
Adenoma Dysplasia Grading of Colorectal Polyps Using Fast Fourier Convolutional ResNet (FFC-ResNet)
(2023-01-01)
Paing, May Phu
;
Pintavirooj, Chuchart
Colorectal polyps are precursor lesions of colorectal cancer; hence, early detection and dysplasia grading of polyps are essential for determining cancer risk, the possibility of developing subsequent polyps, and follow-up recommendations. The significant contribution of this study is the development of an enhanced deep-learning model called Fast Fourier Convolutional ResNet (FFC-ResNet) to classify dysplasia grades of polyps. It is based on the ResNet-50 architecture and uses cross-feature fusion, which combines local features extracted by traditional spatial convolution with global features extracted by Fourier convolution. Due to the compensatory effect between local and global features, the learnability and performance of FFC-ResNet have increased. The proposed FFC-ResNet was developed and tested using UniToPatho, a dataset containing 7000 μm and 800 μm hematoxylin-and-eosin (H&E)-stained colorectal images. And a favorable performance of sensitivity 0.95, specificity 0.93, balance accuracy 0.94, precision 0.95, F1 score 0.95, and AUC 0.99 was obtained using 800 μm polyp patches.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results