KMITL

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

Browse

Search Results

Now showing 1 - 10 of 10

Some of the metrics are blocked by your
consent settings
Mammogram Analysis with YOLO Models on an Affordable Embedded System
(2026-01-01)
Intasam, Anongnat
;
Piyawattanametha, Nicholas
;
Promworn, Yuttachon
;
Jiranantanakorn, Titipon
;
Thawornwanchai, Soonthorn
Background/Objectives: Breast cancer persists as a leading cause of female mortality globally. Mammograms are a key screening tool for early detection, although many resource-limited hospitals lack access to skilled radiologists and advanced diagnostic tools. Deep learning-based computer-aided detection (CAD) systems can assist radiologists by automating lesion detection and classification. This study investigates the performance of various You Only Look Once (YOLO) models and a Hybrid Convolutional-Transformer Architecture (YOLOv5, YOLOv8, YOLOv10, YOLOv11, and Real-Time-DEtection Transformer (RT-DETR)) for detecting mammographic lesions on an affordable embedded system. Methods: We developed a custom web-based annotation tool to enhance mammogram labeling accuracy, using a dataset of 3169 patients from Thailand and expert annotations from three radiologists. Lesions were classified into six categories: Masses Benign (MB), Calcifications Benign (CB), Associated Features Benign (AFB), Masses Malignant (MM), Calcifications Malignant (CM), and Associated Features Malignant (AFM). Results: Our results show that the YOLOv11n model is the optimal choice for the NVIDIA Jetson Nano, achieving an accuracy of 0.86 and an inference speed of 6.16 ± 0.31 frames per second. A comparative analysis with a graphics processing unit (GPU)-powered system revealed that the Jetson Nano achieves comparable detection performance at a fraction of the cost. Conclusions: The current research landscape has not yet integrated advanced YOLO versions for embedded deployment in mammography. This method could facilitate screening in clinics without high-end workstations, demonstrating the feasibility of deploying CAD systems in low-resource environments and underscoring its potential for real-world clinical applications.
Some of the metrics are blocked by your
consent settings
Rapid Detection of Perfluorooctanesulfonic Acid Using Surface-Enhanced Raman Spectroscopy and Deep Learning
(2025-10-07)
Juhong, Aniwat
;
Li, Bo
;
Liu, Yifan
;
Yao, Cheng You
;
Yang, Chia Wei
Per- and polyfluoroalkyl substances (PFAS) are a large group of human-made chemicals that have been widely used in industry and consumer products. Perfluorooctanesulfonic acid (PFOS) is a ubiquitous type of PFAS, which is extremely stable chemicals that have been persistent in the environment for many years. The accumulation of PFOS in the human body can lead to various unfavorable health issues related to the immune, metabolic, and endocrine systems. The conventional PFOS detection method utilizes liquid chromatography coupled with a mass spectroscopy system that typically involves a lengthy and complex procedure. Herein, we propose to develop a low-cost and rapid test approach based on surface-enhanced Raman spectroscopy (SERS) and deep learning for PFOS detection. The gold nanoparticle SERS substrates utilized in this study can significantly enhance the Raman signal of PFOS in solution at a low concentration. PFOS detection and quantification in water using the SERS-based substrate are carried out by measuring Raman peak intensities of PFOS in solution at a range of low concentrations and comparing them to the signal of a blank SERS substrate background. The results show that the SERS substrate can achieve a detection limit as low as 0.0005 ppb. In addition, we propose a demultiplexing deep learning model, which can generate high signal-to-noise ratio (SNR) PFOS spectra from the noisy mixture of PFOS and background Raman spectra. Average cross-correlation and mean absolute error (MAE) are utilized to evaluate the similarity between the demultiplexed and denoised PFOS Raman spectra (output of deep learning) and their ground truths. The proposed model can achieve an encouraging result with high average cross-correlation and low average MAE of 0.9622 ± 0.0667 and 0.0034 ± 0.0024, respectively.
Some of the metrics are blocked by your
consent settings
Monocular depth estimation based on deep learning for intraoperative guidance using surface-enhanced Raman scattering imaging
(2025-02-01)
Juhong, Aniwat
;
Li, Bo
;
Liu, Yifan
;
Yao, Cheng You
;
Yang, Chia Wei
Imaging of surface-enhanced Raman scattering (SERS) nanoparticles (NPs) has been intensively studied for cancer detection due to its high sensitivity, unconstrained low signal-to-noise ratios, and multiplexing detection capability. Furthermore, conjugating SERS NPs with various biomarkers is straightforward, resulting in numerous successful studies on cancer detection and diagnosis. However, Raman spectroscopy only provides spectral data from an imaging area without co-registered anatomic context. This is not practical and suitable for clinical applications. Here, we propose a custom-made Raman spectrometer with computer-vision-based positional tracking and monocular depth estimation using deep learning (DL) for the visualization of 2D and 3D SERS NPs imaging, respectively. In addition, the SERS NPs used in this study (hyaluronic acid-conjugated SERS NPs) showed clear tumor targeting capabilities (target CD44 typically overexpressed in tumors) by an ex vivo experiment and immunohistochemistry. The combination of Raman spectroscopy, image processing, and SERS molecular imaging, therefore, offers a robust and feasible potential for clinical applications.
Some of the metrics are blocked by your
consent settings
Multihead Attention U-Net for Magnetic Particle Imaging–Computed Tomography Image Segmentation
(2024-10-01)
Juhong, Aniwat
;
Li, Bo
;
Liu, Yifan
;
Yang, Chia Wei
;
Yao, Cheng You
Magnetic particle imaging (MPI) is an emerging noninvasive molecular imaging modality with high sensitivity and specificity, exceptional linear quantitative ability, and potential for successful applications in clinical settings. Computed tomography (CT) is typically combined with the MPI image to obtain more anatomical information. Herein, a deep learning-based approach for MPI-CT image segmentation is presented. The dataset utilized in training the proposed deep learning model is obtained from a transgenic mouse model of breast cancer following administration of indocyanine green (ICG)-conjugated superparamagnetic iron oxide nanoworms (NWs-ICG) as the tracer. The NWs-ICG particles progressively accumulate in tumors due to the enhanced permeability and retention (EPR) effect. The proposed deep learning model exploits the advantages of the multihead attention mechanism and the U-Net model to perform segmentation on the MPI-CT images, showing superb results. In addition, the model is characterized with a different number of attention heads to explore the optimal number for our custom MPI-CT dataset.
Some of the metrics are blocked by your
consent settings
Recurrent and convolutional neural networks for sequential multispectral optoacoustic tomography (MSOT) imaging
(2023-11-01)
Juhong, Aniwat
;
Li, Bo
;
Liu, Yifan
;
Yao, Cheng You
;
Yang, Chia Wei
Multispectral optoacoustic tomography (MSOT) is a beneficial technique for diagnosing and analyzing biological samples since it provides meticulous details in anatomy and physiology. However, acquiring high through-plane resolution volumetric MSOT is time-consuming. Here, we propose a deep learning model based on hybrid recurrent and convolutional neural networks to generate sequential cross-sectional images for an MSOT system. This system provides three modalities (MSOT, ultrasound, and optoacoustic imaging of a specific exogenous contrast agent) in a single scan. This study used ICG-conjugated nanoworms particles (NWs-ICG) as the contrast agent. Instead of acquiring seven images with a step size of 0.1 mm, we can receive two images with a step size of 0.6 mm as input for the proposed deep learning model. The deep learning model can generate five other images with a step size of 0.1 mm between these two input images meaning we can reduce acquisition time by approximately 71%.
Some of the metrics are blocked by your
consent settings
Cost-Effective Near Infrared Fluorescence Wide-Field Camera for Breast Tumor Imaging
(2023-08-01)
Juhong, Aniwat
;
Li, Bo
;
Yao, Cheng You
;
Yang, Chia Wei
;
Liu, Kunli
Near-infrared (NIR) fluorescence imaging has long been proven useful for numerous medical applications, especially image-guided surgery since it can deeply penetrate tissue and acquire a fluorescence image from the FDA-approved fluorescence dye known as indocyanine green (ICG). NIR imaging-guided surgery can help enhance outcomes by better defining surgical margins and minimizing operative time. However, most commercial NIR fluorescence imaging systems are bulky and expensive. In this work, we introduce a low-cost and portable NIR fluorescence imaging system. The proposed system exploits an inexpensive complementary metal-oxide-semiconductor (CMOS) based on Nyxel® technology, considerably improving quantum efficiency (QE) in the NIR range. Therefore, it can acquire a NIR fluorescence image with outstanding quality. Furthermore, we conjugated ICG with nanoworm particles (NWs-ICG) since nanoparticle-conjugated dye can increase accumulation in a tumor. We show impressive applications of our system for ex vivo tissues and in vivo image-guided surgery using NWs-ICG as a contrast agent for breast tumors from transgenic mice.
Some of the metrics are blocked by your
consent settings
Super-resolution and segmentation deep learning for breast cancer histopathology image analysis
(2023-01-01)
Juhong, Aniwat
;
Li, Bo
;
Yao, Cheng You
;
Yang, Chia Wei
;
Agnew, Dalen W.
Traditionally, a high-performance microscope with a large numerical aperture is required to acquire high-resolution images. However, the images’ size is typically tremendous. Therefore, they are not conveniently managed and transferred across a computer network or stored in a limited computer storage system. As a result, image compression is commonly used to reduce image size resulting in poor image resolution. Here, we demonstrate custom convolution neural networks (CNNs) for both super-resolution image enhancement from low-resolution images and characterization of both cells and nuclei from hematoxylin and eosin (H&E) stained breast cancer histopathological images by using a combination of generator and discriminator networks so-called super-resolution generative adversarial network-based on aggregated residual transformation (SRGAN-ResNeXt) to facilitate cancer diagnosis in low resource settings. The results provide high enhancement in image quality where the peak signal-to-noise ratio and structural similarity of our network results are over 30 dB and 0.93, respectively. The derived performance is superior to the results obtained from both the bicubic interpolation and the well-known SRGAN deep-learning methods. In addition, another custom CNN is used to perform image segmentation from the generated high-resolution breast cancer images derived with our model with an average Intersection over Union of 0.869 and an average dice similarity coefficient of 0.893 for the H&E image segmentation results. Finally, we propose the jointly trained SRGAN-ResNeXt and Inception U-net Models, which applied the weights from the individually trained SRGAN-ResNeXt and inception U-net models as the pre-trained weights for transfer learning. The jointly trained model’s results are progressively improved and promising. We anticipate these custom CNNs can help resolve the inaccessibility of advanced microscopes or whole slide imaging (WSI) systems to acquire high-resolution images from low-performance microscopes located in remote-constraint settings.
Some of the metrics are blocked by your
consent settings
Optimizing the Hyperparameter Tuning of YOLOv5 for Breast Cancer Detection
(2023-01-01)
Intasam, Anongnat
;
Promworn, Yuttachon
;
Juhong, Aniwat
;
Thanasitthichai, Somchai
;
Khwayotha, Sirihattaya
This research aims to find the best use optimizer for the task while reducing training time. We optimized the YOLOv5s model and focused on three optimizers, including the Stochastic Gradient Descent (SGD) optimizer, Adaptive Moment Estimation (Adam) optimizer, and Adam with Weight Decay Regularization (AdamW) optimizer. This research utilized 1,471 mammogram images from National Cancer Institute and Udonthani Cancer Hospital, Thailand. A dataset of mammograms was labeled into six classes, including Masses Benign, Masses Malignant, Calcifications Benign, Calcifications Malignant, Associated Features Benign, and Associated Features Malignant, to classify the results accurately. We found that the SGD optimizer outperformed the others, with a mean average precision (mAP) of 0.91, a precision of 0.92, a recall of 0.85, and the shortest training time of about 5.453 hr.
Some of the metrics are blocked by your
consent settings
Medical Drone Managing System for Automated External Defibrillator Delivery Service
(2022-04-01)
Purahong, Boonchana
;
Anuwongpinit, Thanavit
;
Juhong, Aniwat
;
Kanjanasurat, Isoon
;
Pintaviooj, Chuchart
One of the common causes of a heart attack is fibrillation, a condition that causes an irregular and often abnormally fast heart rate. There is scientific evidence that the survival rate of sudden cardiac arrest patients who are rescued with cardiopulmonary resuscitation (CPR) and with the use of an automated external defibrillator (AED) is significantly increased. Despite the recommendation that automated external defibrillators should be installed in the workplace, along with a proper management system and training for employees on how to use the device, less than 70% of non-residential areas have an AED installed. The situation is even worse in residential areas, with less than 30% having an AED installed. This research concerns the development of a medical drone managing system that can deliver an AED in case of emergency. An application was developed that can be installed on the mobile phone and/or tablet of the patient or the accompanying person. In the event of a heart attack, the patient or the accompanying person can call a medical drone by sending coordinates to the drone station and a notification to medical staff. The drone station administrator can respond by sending the drone, which automatically lands at the patient’s location. After being tested in a simulation situation, the operational field test yielded satisfactory results. The medical drone can land within 1.5 meters of the destination. The designed AED drone can be used not only to deliver AEDs, but also first aid kits and prescribed drugs suitable for medical care. Such a system is especially useful in the current context of the COVID‐19 pandemic.
Some of the metrics are blocked by your
consent settings
Design and Development of an Assistive System Based on Eye Tracking
(2022-02-01)
Paing, May Phu
;
Juhong, Aniwat
;
Pintavirooj, Chuchart
This research concerns the design and development of an assistive system based on eye tracking, which can be used to improve the quality of life of disabled patients. With the use of their eye movement, whose function is not affected by their illness, patients are capable of communicating with and sending notifications to caretakers, controlling various appliances, including wheelchairs. The designed system is divided into two subsystems: Stationary and mobile assistive systems. Both systems provide a graphic user interface (GUI) that is used to link the eye tracker with the appliance control. There are six GUI pages for the stationary assistive system and seven for the mobile assistive system. GUI pages for the stationary assistive system include the home page, smart appliance page, eye-controlled television page, eye-controlled air conditional page, i-speak page and entertainment page. GUI pages for the mobile assistive system are similar to the GUI pages for the stationary assistive system, with the additional eye-controlled wheelchair page. To provide hand-free secure access, an authentication based on facial landmarks is developed. The operational test of the proposed assistive system provides successful and promising results.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results