In this paper we present a new method for automated recognition of 12 microalgae that are most commonly found in water resources Thailand. order to handle some difficulties encountered our problem such as unclear algae boundary and noisy background, proposed segmenting bodies from an image background computing texture descriptors blurry object. Feature combination approach is applied variation shapes the same genus. Sequential Minimal Optimization (SMO) used classifier. An experimental...

Most existing studies on credit scoring adapted a concept of classifier ensemble for solving an imbalanced dataset. They apply resampling methods to generate multiple training subsets constructing base classifiers. However, this approach leads several problems that degrade the classification performance, such as information loss, model overfitting, and computational cost. Thus, we propose novel developing based cost-sensitive neural network, called Cost-sensitive Neural Network Ensemble...

In this paper we propose an algorithm for combining multiple image segmentations to achieve a final improved segmentation. contrast previous works consider the most general class of segmentation combination, i.e. each input has arbitrary number regions. Our approach is based on random walker which able provide high-quality starting from manually specified seeds. We automatically generate such seeds ensemble. An information-theoretic optimality criterion proposed determine The experimental...

In this paper we present the adaptation of a random walker algorithm for combination image segmentations to work with clustering problems. order achieve it, pre-process ensemble clusterings generate its graph representation. We show experimentally that very small neighborhood will produce similar results if compared larger choices. This fact alone improves computational time needed final consensual clustering. also an experimental comparison between our against other based and well known...

We have developed a portable electronic nose (E-nose) based on eight metal oxide gas sensors for classification and prediction of meat freshness. In this study, the E-nose was applied to predict chicken freshness during different storage days. Principal component analysis (PCA) artificial neural network (ANN) were used analyze experiment data. The PCA method can classify related ANN result shows good agreement with result. correct rate in is 97.92%. From results, it indicates that well owns...

<span lang="EN-US">In the process of bankruptcy prediction models, a class imbalanced problem has occurred which limits performance models. Most prior research addressed by applying resampling methods such as synthetic minority oversampling technique (SMOTE). However, lead to other issues, e.g., increasing noisy data and training time during process. To improve model, we propose cost-sensitive extreme gradient boosting (CS-XGB) address without requiring any method. The proposed...

<span>Several credit-scoring models have been developed using ensemble classifiers in order to improve the accuracy of assessment. However, among models, little consideration has focused on hyper-parameters tuning base learners, although these are crucial constructing models. This study proposes an improved credit scoring model based extreme gradient boosting (XGB) classifier Bayesian optimization (XGB-BO). The comprises two steps. Firstly, data pre-processing is utilized handle...

Due to rapid advance of computer vision technology, assisted image analysis starts play an important role in several areas including aquaculture. In recent years vision-based methods have been applied many major operations, e.g. automated fish counting, inspection, and measurement. this paper we address a problem overlapping objects population that frequently occurs when under investigation are allowed move freely during operations. We proposed new skeleton reconstruction algorithm for...

The region-based segmentation paradigm is a well known technique for image segmentation. In the first part of this work robustness algorithms studied. It shown that within small parameter range, which leads to good results in majority cases, bad may occur. fact, such local instability problem methods and reasons its occurrence are discussed. second work, an ensemble solution based on median concept proposed. Two variants, set generalized median, presented experimentally compared. Extensive...

Khao Dawk Mali 105 (KDML105), internationally known as "Jasmine Rice", is one of the most famous and major commercial rice in Thailand. Physical appearance polished grain key factors that influences rice's price. One traits a degree chalkiness. Rice breeding scientists thus make great effort reducing chalkiness order to meet market quality match consumer preference. The routine task process visually inspection level rice. Since human visual slow, subjective, not consistent over long period,...

Generally, the purpose of saliency detection models for object and fixation prediction is complementary. Saliency aim to discover as much possible true positive, while intend generate few false positive. In this work, we attempt combine their strength together. We accomplish by, firstly, replacing high-level features that frequently used in a model with our new location map order make more general. Secondly, train human eye tracking data correspond well (without use top-down attention)....

Cluster ensemble has emerged as a powerful technique for improving robustness, stability, and accuracy of clustering solutions. In this paper we present novel use cluster to handle another most difficult problem in data - model order selection. Each component is viewed an expert domain building the case-based reasoning. Our proposed method simple fast, but effective. Three simulations with different state-of-the-art segmentation algorithms are presented illustrate efficacy approach. We...

Malaria is a significant global health issue, with 241 million people infected and resulting in 627,000 deaths 2020, officially reported by the World Health organization (WHO). In addition, during Covid-19 pandemic, 47,000 died because of reluctance to receive treatment. Thailand, still spreads distant communities where restrictions are place for military deployments due high risk infection. Therefore, 8,000 or so personnel who deploy on missions close country's borders actively monitored...