KMITL

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

Browse

Search Results

Now showing 1 - 10 of 31

Some of the metrics are blocked by your
consent settings
Identification and quantification of quality of intact durian fruits using NIR spectroscopy
(2026-01-01)
Pitak, Lakkana
;
Ditcharoen, Sirirak
;
Maraphum, Kanvisit
;
Khamwan, Buathip
;
Warorost, Nithithada
Quality classification of durian fruits is based on the dry matter (DM) content of the pulp. According to Thai agricultural standards, durian fruit (Monthong variety) must contain at least 32% DM. This study aimed to develop a classification model for assessing durian quality based on DM content, categorizing fruits as either “rejected” (DM < 32%) or “accepted” (DM ≥ 32%). Near-infrared (NIR) spectra were collected as the durian fruits moved along a conveyor belt. The models were developed using two spectral ranges: short-wavelength near-infrared (SWNIR; 4501000 nm) and long-wavelength near-infrared (LWNIR; 8601750 nm). Owing to the imbalance in the dataset between the two classes, the data were adjusted using the synthetic minority oversampling technique to create a balanced dataset. Prediction models were built using different spectral preprocessing methods and algorithms. For the LWNIR range, the models constructed using LDA, SVM, KNN, and SDA achieved accuracies of 95%, 90%, 93%, and 93%, respectively, for the test set. The SWNIR models, developed using the same algorithms, achieved accuracies of 90%, 88%, 90%, and 90%, respectively, for the test set. PLS-regression was used to predict the DM content from both LWNIR and SWNIR data. With the 2nd derivative preprocessing method, the models achieved R² values of 0.89 and 0.79, SEP values of 5% and 6.89%, and RPD values of 2.29 and 1.66, respectively. The wavelength range significantly influenced the model performance, whereas spectral pretreatment had a minor effect on the model's predictive ability. Overall, NIR spectroscopy demonstrated the potential for nondestructive quality grading of whole durian fruits. This work is the first to establish real-time, in-line models for durian grading based on DM content, advancing beyond the previous destructive method. The findings demonstrate the feasibility of automated, nondestructive, and objective quality assessment, supporting industrial automation, precision agriculture, and export quality assurance.
Some of the metrics are blocked by your
consent settings
Rapid detection of potassium sorbate in coconut water using near infrared hyperspectral imaging
(2026-01-01)
Tantinantrakun, Achiraya
;
Kumpa, Benjaporn
;
Ainkast, Pranpriya
;
Thompson, Anthony Keith
;
Teerachaichayut, Sontisuk
Potassium sorbate may be illegally added to fresh coconut water in order to prolong its marketable life, but this adulteration may not be identified on the product label. The aim of this research was therefore to evaluate if samples of fresh coconut water that had been adulterated with measured amounts of potassium sorbate could be detected by near infrared hyperspectral imaging (NIR-HSI). Samples of coconut water with different potassium sorbate concentrations (N = 100) and pure coconut water samples (N = 100) were used in this study with their averaged spectral data used as independent variables. The smoothing spectral pretreatment gave the highest classification accuracy of 98.48% by partial least squares discriminant analysis (PLS-DA). While support vector machine regression (SVMR) with spectral pretreatment, using the 1st derivative combined with multiplicative scatter correction (MSC), achieved the optimum condition for developing the calibration model for determining potassium sorbate concentration with the correlation coefficient of prediction (R<inf>p</inf>) of 0.818 and the root mean square error of prediction (RMSEP) of 327.86 ppm. The results showed that NIR-HSI was able to be used as a fast, reliable, economic and environmentally friendly method of detecting potassium sorbate addition to coconut water.
Some of the metrics are blocked by your
consent settings
Performance of deep learning models for the classification and object detection of different oral white lesions using photographic images
(2025-12-01)
Khovidhunkit, Siribang on Piboonniyom
;
Phosri, Kunchidsong
;
Thanathornwong, Bhornsawan
;
Rungraungrayabkul, Dulyapong
;
Poomrittigul, Suvit
Computer vision adjunctive technology for oral lesion diagnoses has been developed to detect and identify Oral Potentially Malignant Disorders (OPMDs) and non-OPMDs. The early detection of OPMDs can reduce the risk of oral cancer development, improving the survival rate of the patients. This study aims to evaluate the computer vision technique in the white oral lesion domain within the scope of photographic images. Deep learning techniques for the classification of Convolution Neural Networks (CNNs) and transformer neural networks, and one-stage models of YOLOv7 and YOLOv8 were utilized to classify and detect five classes of OPMDs and non-OPMDs oral white lesions including oral leukoplakia, oral lichen planus, pseudomembranous candidiasis, oral ulcers covered with pseudomembrane and other white benign oral lesions. From the evaluation results of classification, the IFormerBase model achieves overperformance compared to CNN models with accuracy, precision, and F1 score of more than 80% on the test set. The best model for object detection is YOLOv7 with 84.5% mean Average Precision (mAP) at Intersection over Union (IoU) threshold of 0.3 and 74.5% at IoU of 0.5 on the test set. Object detection results reveal promising automatic oral lesion identification, which can be further developed to enhance the lesion screening system.
Some of the metrics are blocked by your
consent settings
A Self-Adaptive Weights for K-Means Classification Algorithm
(2025-09-01)
Chenghu, Cui
;
Thammano, Arit
This paper presents an improved K-means clustering algorithm that addresses the traditional algorithm’s sensitivity to outlier and susceptibility to local optima by introducing an adaptive weight adjustment mechanism. It employs an exponential decay function to dynamically reduce the feature weights of outlier data points, effectively suppressing outliers while preserving the structure of the normal data. The proposed method retains the computational efficiency of standard K-means. Key contributions include: (a) A novel distance-based weighting strategy that progressively reduces the influence of noisy points, mitigating the impact of outliers on clustering performance. (b) An innovative form of "local dimensionality reduction" for outlier points via weight decay, which interferes only with the feature space of noisy regions while preserving the global topological structure of clean data. Extensive experiments on three benchmark datasets Iris (4-dimensional, balanced classes), Wine (13-dimensional, correlated features), and Wisconsin Breast Cancer Diagnosis (30-dimensional, imbalanced data) demonstrate the effectiveness of the approach. Compared to standard K-means, the proposed algorithm achieves accuracy improvements of 7.47% on Iris, 13.89% on Wine, and 19% on WBCD. This adaptive strategy offers a practical and efficient solution for clustering in noisy, high-dimensional environments, without the added complexity of mixture models.
Some of the metrics are blocked by your
consent settings
Digital holography with deep learning for algae identification and classification
(2024-01-01)
Ruttanasirawit, Chinnaphat
;
Plaipichit, Suwan
;
Thongsuwan, Setthanun
;
Thonglim, Pachara
;
Phunpruch, Saranya
Recently, the characterization of marine objects, populations and biophysical interactions have become crucial within the research community. In this study, we leverage digital holographic imaging systems and deep learning networks to classify three distinct types of micro-algae: Chlamydomonas, Scenedesmus armatus, and Scenedesmus_sp-L. We employed reconstructed digital holographic images and deep learning to identify the results from both approaches. The integration of holographic imaging holds promises in replacing expensive characterization systems like AFM, x-ray diffraction, and Raman spectroscopy, offering a more costeffective solution. In our system, we utilize in-line microscopic digital holographic imaging to record and reconstruct images of the algae specimens. An essential advantage of holographic techniques is that they do not require intact samples of the specimens for effective object identification. To further enhance the process, we combined deep learning algorithms with holographic imaging, capitalizing on the advanced computers. This combination enables highly effective characterizing and classification of different types of algae. These innovative approaches pave the way for exciting advancement in marine research and monitoring.
Some of the metrics are blocked by your
consent settings
DEEP LEARNING DEVELOPMENTAL MODEL FOR EMBRYO GRADING IN THAI POPULATION
(2024-01-01)
Wiriyasirivaj, Budsaba
;
Limkiatsataporn, Sawit
;
Pukinghin, Apisit
;
Kuekulkomain, Phatrapron
;
Promrungrueng, Pornprom
In light of the growing challenges associated with infertility, an increasing number of researchers are resorting to assisted reproductive technologies such as In Vitro Fertilization (IVF). Embryo grading is a crucial step in the IVF process that requires embryologists’ expertise. However, their limited availability has led to the exploration of technological alternatives. This study aims to use deep learning for human embryo grading, with models specifically designed for the dataset in Thailand at Vajira Hospital. The process of IVF at Vajira Hospital presents its own set of challenges since its embryo classification extends beyond the Istanbul consensus. Furthermore, classes that occur infrequently are removed and classes with similarities are merged. We apply transfer learning to pretrained deep learning models like VGG19, VGG16, ResNet152, Resnet101, Xception, InceptionV3, and EfficientNet, to create a system capable of accurately classifying embryo quality scores. Experimental results from the embryo dataset collected from Vajira Hospital in Thailand demonstrate the proposed classifier’s accuracy, precision, recall, f1-score, and AUC superiority. This research contributes to the field of IVF in Thailand by potentially reducing human errors and addressing the demand for skilled embryologists.
Some of the metrics are blocked by your
consent settings
Partial Discharge Classification with Transformer Neural Networks
(2024-01-01)
Cheypoca, Thepjit
;
Promphanich, Wiboon
;
Thway, Aung Ye
;
Hankae, Angelina Phimpissada
;
Jeenmuang, Siwakorn
This paper introduces an approach with the Transformer Neural Networks model for partial discharge patterns classification, that consists of corona discharge, internal discharge and surface discharge. The PD measuring circuit suggested in IEC 60270:2000 is used to record Partial discharge signals. Independent parameters such as phase and charge of PD patterns were recorded. The phase value will be encoded into the charge array and Transformer Neural Network is constructed using Positional Embedding and Transformer Encoder Layer. 80% of the recorded data will be used as a training data and 20% recorded data was used for testing of the classification models. Impacts of neuron numbers and network architecture on the PD classification performance will be observed
Some of the metrics are blocked by your
consent settings
Subcellular Protein Patterns Classification Using Extreme Gradient Boosting with Deep Transfer Learning as Feature Extractor
(2024-01-01)
Phankokkruad, Manop
;
Wacharawichanant, Sirirat
Proteins are essential structural and functional components of human cells. Understanding and identifying proteins can provide valuable insights into their structure, function and role in human body. Subcellular proteins provide the expression that characterizes the many proteins and their conditions across cell types. This work proposed a classification model for subcellular protein patterns using XGBoost with transfer learning of CNN as the feature extractor. In the model training process, we used ResNet50, VGG16, Xception, and MobileNet as the pre-trained models based on the transfer learning technique to extract different features. The proposed model was used to classify subcellular proteins into 28 patterns. The XGBoost with ResNet50, VGG16, MobileNet, and Xception model achieved an accuracy level of 92.20% 92.77%, 91.63%, and 91.44%, respectively. The XGBoost with ResNet50, VGG16, MobileNet, and Xception model obtained an F1 score of 0.9179, 0.9233, 0.9131, and 0.9152, respectively. Considering the F1 score, All XGBoost with transfer learning of CNN models gave a high score. Therefore, all evaluation parameters clearly demonstrate the high performance of the subcellular protein pattern classification model.
Some of the metrics are blocked by your
consent settings
Total soluble solids, dry matter content prediction and maturity stage classification of durian fruit using long-wavelength NIR reflectance
(2023-12-01)
Saenphon, Chirawan
;
Ditcharoen, Sirirak
;
Malai, Chayuttapong
;
Saengprachatanarug, Khwantri
;
Wongpichet, Seree
The DM and TSS of durian pulp moving on a conveyor belt were measured for their rapid and non-destructive qualities based on quantitative and qualitative measurements. The calibration set and prediction set equaled 209 and 69 pulps, respectively. The quantitative test compared the performance of PLS regression for DM and TSS prediction developed from full wavelength (860–1754 nm) and a few significant variables using SPA, GA, and VIP methods. The qualitative test identified the possibility of maturity stage classification by comparing three supervised machine learning classifiers, namely SVM, random forest (RF), and LDA. Effective models for DM and TSS prediction were developed from second derivatives spectra combined with the GA method, exhibiting r<sup>2</sup>, SEP, and RPD of 0.85, 4.50%, and 2.64, respectively for DM, and 0.66, 5.15%, and 1.60, respectively, for TSS. The model classifying samples into two distinct groups, namely “reject” and “pass,” utilizing the LDA algorithm, exhibited an impressive accuracy rate of 94.20%, making it a suitable choice for quality assurance purposes. This result indicates that the few effective variables were more efficient than full wavelength and improved model accuracy with greater model stability. Enhancing the classification model could involve data sample balancing in each group, leading to further improvements.
Some of the metrics are blocked by your
consent settings
Rapid and accurate classification of Aspergillus ochraceous contamination in Robusta green coffee bean through near-infrared spectral analysis using machine learning
(2023-03-01)
Ruttanadech, Nuttapong
;
Phetpan, Kittisak
;
Srisang, Naruebodee
;
Srisang, Siriwan
;
Chungcharoen, Thatchapol
Near-infrared (NIR) spectral-based classification of Aspergillus ochraceous contamination in the Robusta green coffee bean was investigated. Six different learning algorithms, including linear discriminant analysis (LDA), support vector machine (SVM), k-nearest neighbors (KNN), decision tree (Tree), Naive Bayes (NB), and quadratic discriminant analysis (QDA), were applied for the investigating purpose. Four classes of fungal contamination on coffee beans, non-fungal contaminated beans on day 1 and day 3 (NCB-D1 and NCB-D3) and fungal contaminated beans on day 1 and day 3 (CB-D1 and CB-D3), were set for the classification intention. Based on the 6 learning algorithms, the Tree approach was optimal, displaying a training accuracy of 97.5%. As proven by the testing dataset, the classification accuracy of the Tree was also at 97.5%. With this number, the Tree could correctly classify 100% between the contaminated and non-contaminated coffee beans. These findings exhibit the potential of the NIR spectroscopy accompanied by machine learning for the early detection of fungal contamination in green coffee beans.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results