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
An artificial intelligence model for the diagnosis of otitis media with effusion in children
(2026-01-01)
Ungkanont, Kitirat
;
Udomchaiporn, Akadej
;
Sriphoonga, Nopavit
;
Wannarong, Thanakrit
;
Rugsujrit, Thaweewat
Background: The diagnosis of otitis media with effusion (OME) requires substantial training and experience in otoscopic examination of children. Objective: This study developed an artificial intelligence (AI) model to predict OME diagnosis in children. Methods: The source data were images of pediatric patients’ tympanic membranes obtained by otoendoscopy. A convolutional neural network was used in machine learning. The diagnostic features of the tympanic membrane, as labelled by the experts, and the surgical findings served as the ground truth. InceptionV4 built the final model. The model was trained using the Adaptive Moment Estimation optimizer with an initial learning rate of 0.0001 and a total duration of 100 epochs. The batch size was 32. The Categorical Cross-Entropy loss function was employed for the internal validation. The outcome was to distinguish between OME and normal tympanic membrane. A confusion matrix was used to assess the model’s performance. The model was tested for agreement with otolaryngologists and implemented as a web application. Results: The initial sample size was 320 pictures. For OME, the model achieved an accuracy of 94.7% (95% CI 0.88, 1). The F1 score was 96% (95% CI 0.89, 1), and the area under the receiver operating characteristic curve was 0.98 (95% CI 0.93, 1). The kappa agreement between AI and experienced otolaryngologists was 0.627 (p < 0.001). Conclusion: An AI diagnostic model for otitis media with effusion had good accuracy and moderate agreement with otolaryngologists. The model should be helpful for preliminary diagnosis, telemedicine, or educational purposes.
Some of the metrics are blocked by your
consent settings
Hybrid Deep Learning Framework for Accurate Surface Defect Detection Using Autoencoder and CNNs
(2026-01-01)
Craypo, Niphat
;
Banjongkan, Anupong
;
Hanskunatai, Anantaporn
Surface defect detection is paramount in industrial quality control. Conventional methods, which often rely on human inspection or manually engineered statistical models, frequently fail to accurately detect and classify defects, particularly on complex surfaces or those with intricate features. Human inspection is inherently inconsistent and prone to errors due to fatigue, while traditional machine vision systems often lack the sensitivity to clearly identify small or low-contrast defects. This paper proposes a hybrid deep learning framework, termed Autoencoders-Convolutional Neural Networks (AE-CNNs), DefectNet, for surface defect classification, AE, and CNNs to enhance both accuracy and efficiency. The AE is employed to extract and compress reliable features from surface images into latent representations, which are subsequently classified by a CNN enhanced through transfer learning using InceptionV3. The CNN is fine-tuned from a pretrained model with customized fully connected layers to adapt to specific defect characteristics, while the AE is trained exclusively on non-defective images. The encoded features produced by the AE serve as the input to the CNN. The proposed model is evaluated on standard benchmark datasets comprising diverse surface defect types and compared against Anomaly Detection with Autoencoder (ADA), Visual Geometry Group (VGG16), Inception-based Convolutional Neural Network Long Short-Term Memory (In-CNNLSTM), and DTL_Inception_v3. Experimental results demonstrate the superior performance of the proposed method, achieving classification accuracies ranging from 85.60% to 100% across five datasets, including a perfect 100% accuracy on the glass bottle neck dataset.
Some of the metrics are blocked by your
consent settings
Glioma Brain Tumor Classification Using Convolution Neural Network and Majority Voting
(2025-01-01)
Pilaoon, Pongsak
;
Hamamoto, Kazuhiko
;
Maneerat, Noppadol
Glioma brain tumors are malignant diseases for which early detection and instant treatment will increase the survival rate. Several studies have reported the efficiency of deep learning convolutional neural networks (CNN) in diagnosing brain tumors using magnetic resonance imaging (MRI). In this study, we investigated the potential of state-of-the-art classifiers to achieve the highest accuracy in the detection of brain tumors using MRI. For this purpose, we introduced a comparative study of eight state-of-the-art classifiers. The methodology comprised three different approaches: 1) an imbalanced dataset, 2) a balanced dataset using image augmentation, and 3) ensemble learning using the best of the top five models for majority hard and soft voting. The dataset comprised converted MRI data from the repository of molecular brain neoplasia data (REMBRANDT) and brain tumor segmentation 2021 (BraTS) databases. An increasing number of MRI images and datasets has prevented overfitting. Initially, a preprocessing stage morphological operation and contrast-limited adaptive histogram equalization (CLAHE) algorithms were used to remove skeletons and artifacts and optimize the image contrast for readiness classification. The stochastic gradient descent with the momentum algorithm option was used to train the network. The trained model was used to predict the testing dataset, and the results from each pretrained network were evaluated. The experimental results demonstrated that the prediction accuracy of the trained network was significantly improved using a balanced training dataset. The discriminative image region used to interpret the predicted result using the gradient-weighted class activation mapping (Grad-CAM) algorithm was proposed in the final stage for trustworthiness. The experimental results showed that the best approach was inceptionV3 with a balanced dataset. The accuracy, sensitivity, specificity, and area under the curve were 99.73%, 99.61%, 100%, and 1.00, respectively.
Some of the metrics are blocked by your
consent settings
Minimizing Model Size of CNN-Based Vehicle Make Recognition for Frontal Vehicle Images
(2025-01-01)
Puisamlee, Wiput
;
Chawuthai, Rathachai
Vehicle Make Model Recognition (VMMR) is commonly used in Intelligent Transportation Systems (ITS), free-flow image-based toll systems, and enforcement systems. These systems must analyze and process vehicle front images for use as evidence. Convolutional Neural Networks (CNN) are widely used for image classification and VMMR problems. Complex model structures and more internal parameters are needed to improve classification accuracy with many classes. Issues included larger models and longer processing times. The goal of this work is to study and create a smaller CNN model that can be used on devices with limited resources, like embedded computers and embedded computer cameras, to figure out what kind of car it is from a front view picture. Real free-flow toll systems were used to train a CNN model that recognized vehicle makes with 99% accuracy. The model is smaller than VGG16, InceptionV3, Yolo11m-cls, and ResNet50 and has over 90% accuracy. It reduced parameters by 69.95% and developed the CTv1 model to achieve an F1 score 2.06% higher than InceptionV3, the best. The model was tested on a Raspberry Pi 3 Model B, processing images in 1 second and using 25 mWh. The compact version of the proposed model also adjusts the Padding and Stride of the Convolutional Layer and reduces the CNN model size using Depth-wise Separable Convolutional and 1 × 1 Convolutional Dimension Reduction (Bottleneck) methods to test vehicle make recognition accuracy, training time, processing time, and model size.
Some of the metrics are blocked by your
consent settings
Fault Detection in Transmission Lines Using CNN
(2024-01-01)
Kanwal, Shazia
;
Jiriwibhakorn, Somchat
Transmission line faults pose a significant risk to power systems, potentially leading to widespread outages. Detecting these faults using advanced algorithms is crucial for preventing major disruptions in power supply. In this paper, we work on a fault detection technique for the IEEE 9-bus system based on deep learning. By training a Convolutional Neural Network (CNN) on features extracted from both normal and faulty conditions, we achieve an accuracy of 86%. This high accuracy underscores the potential of CNNs for real-world implementation in fault detection systems. The robustness of the CNN approach suggests its viability for deployment in complex, real-time systems, offering improved reliability and resilience against transmission line faults. Additionally, utilizing deep learning techniques opens avenues for further refinement and optimization of fault detection strategies in the power grid.
Some of the metrics are blocked by your
consent settings
CNN–RNN Network Integration for the Diagnosis of COVID-19 Using Chest X-ray and CT Images
(2023-02-01)
Kanjanasurat, Isoon
;
Tenghongsakul, Kasi
;
Purahong, Boonchana
;
Lasakul, Attasit
The 2019 coronavirus disease (COVID-19) has rapidly spread across the globe. It is crucial to identify positive cases as rapidly as humanely possible to provide appropriate treatment for patients and prevent the pandemic from spreading further. Both chest X-ray and computed tomography (CT) images are capable of accurately diagnosing COVID-19. To distinguish lung illnesses (i.e., COVID-19 and pneumonia) from normal cases using chest X-ray and CT images, we combined convolutional neural network (CNN) and recurrent neural network (RNN) models by replacing the fully connected layers of CNN with a version of RNN. In this framework, the attributes of CNNs were utilized to extract features and those of RNNs to calculate dependencies and classification base on extracted features. CNN models VGG19, ResNet152V2, and DenseNet121 were combined with long short-term memory (LSTM) and gated recurrent unit (GRU) RNN models, which are convenient to develop because these networks are all available as features on many platforms. The proposed method is evaluated using a large dataset totaling 16,210 X-ray and CT images (5252 COVID-19 images, 6154 pneumonia images, and 4804 normal images) were taken from several databases, which had various image sizes, brightness levels, and viewing angles. Their image quality was enhanced via normalization, gamma correction, and contrast-limited adaptive histogram equalization. The ResNet152V2 with GRU model achieved the best architecture with an accuracy of 93.37%, an F1 score of 93.54%, a precision of 93.73%, and a recall of 93.47%. From the experimental results, the proposed method is highly effective in distinguishing lung diseases. Furthermore, both CT and X-ray images can be used as input for classification, allowing for the rapid and easy detection of COVID-19.
Some of the metrics are blocked by your
consent settings
Banana Plant Nutrient Deficiencies Identification using Deep Learning
(2023-01-01)
Han, Kadipa Aung Myo
;
Maneerat, Noppadol
;
Sepsirisuk, Kasemsuk
;
Hamamoto, Kazuhiko
This paper presents nutrient deficiency multi-class classification in banana plant data sets using a deep convolutional neural network. In this paper, healthy and eight nutrient deficiency classes were studied. The performance was evaluated in different situations of two public data sets. The proposed method can provide sensitivity and specificity in Raw Images, Raw Images with combination, Augmented Images, and Augmented Images with the combination. Furthermore, nearly 88% of the F1-score was outperformed.
Some of the metrics are blocked by your
consent settings
Traffic Light and Crosswalk Detection and Localization Using Vehicular Camera
(2022-01-01)
Wangsiripitak, Somkiat
;
Hano, Keisuke
;
Kuchii, Shigeru
An improved convolutional neural network model for traffic light and crosswalk detection and localization using visual information from a vehicular camera is proposed. Yolov4 darknet and its pretrained model are used in transfer learning using our datasets of traffic lights and crosswalks; the trained model is supposed to be used for red-light running detection of the preceding vehicle. Experimental results, compared to the result of the pretrained model learned only from the Microsoft COCO dataset, showed an improved performance of traffic light detection on our test images which were taken under various lighting conditions and interferences; 36.91% higher recall and 39.21% less false positive rate. The crosswalk, which is incapable of detection in the COCO model, could be detected with 93.37% recall and 7.74% false-positive rate.
Some of the metrics are blocked by your
consent settings
Automatic Lymph Node Classification with Convolutional Neural Network
(2022-01-01)
Uthatham, Ason
;
Yodrabum, Nutcha
;
Sinmaroeng, Chanya
;
Titijaroonroj, Taravichet
Manual lymph node classification is a tedious and time-consuming task. It requires a histopathologist to discriminate a lymph node from other look-alike kinds of tissues. The lymph node is easily misunderstood with other tissues because its shape and color might be similar to the others tissue around it. To automate this task, we present an automatic lymph node classification with convolutional neural network (CNN). In addition, we compared eight existing CNNs to ensure that we discover the best architecture for discriminating lymph node. DenseNet architecture provided the highest performance among AlexNet, VGG, GoogLeNet, ResNet, SqueezeNet, MobileNet, and EfficientNet, the highest accuracy at 0.994 and an F1score of 0.996. DenseNet accomplished the highest performance from two advantages: (i) fewer parameters and (ii) Dense connectivity.
Some of the metrics are blocked by your
consent settings
Musical Key Classification Using Convolutional Neural Network Based on Extended Constant-Q Chromagram
(2022-01-01)
Chivapreecha, Sorawat
;
Sinjanakhom, Tantep
;
Trirat, Akraphon
In the field of music information retrieval, musical key classification is one of the challenges. This paper illustrates the advantages of the proposed system with relevant experimental results, starting with diverse audio datasets for feature extraction used for training and testing a classification model which is based on a convolutional neural network (CNN). The goal is to develop a feature that can improve the neural network's performance. To compare the effect of input features on efficiency, a basic CNN is trained from the ground up and utilized as an image classification tool. The Chromagram-24, an augmented version of the input chroma feature, is proposed to improve the accuracy of musical key detection. In terms of weighted score, the model using Chromagram-24 as an input feature outperforms the model trained using a conventional 12-dimensional chromagram by 12.77% and achieves the highest score of 85.63% when classifying full-length songs. Chromagrams are generated using audio excerpts ranging in length from 15 to 60 seconds for local key estimation, whereas, for global key estimation, a full-length audio set is used. The results indicate that, given the different lengths of training audio input, executing the model using a chromagram of a 60-second audio excerpt yields the best results.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results