In this paper, we introduce a novel method to detect a carried object seen from a stationary camera using human body silhouette feature information. We use star skeletonization technique with the adaptive centroid point to extract human feature. The carried object is classified using time series of motions of the extracted skeleton limbs. The boundary of the carried object is figured from carried object's track points and adjacent sink curves of contour. The method is able to detect and track carried object such as a luggage and a backpack. Moreover, we also achieve the detection of leaving luggage event. We perform experiments using some data from TRECVID dataset and manually captured data.

Lung cancer is a deadly disease with a high mortality rate. Endobronchial ultrasonography (EBUS) is one of the methods for detecting pulmonary lesions. Computer-aided diagnosis of pulmonary lesions from images can help radiologists to classify lesions; however, most of the existing methods need a large volume of data to give good results. Thus, this paper proposes a novel pulmonary lesion classification framework for EBUS images that works well with small datasets. The proposed framework integrates the statistical results from three classification models using the weighted ensemble classification. The three classification models include the radiomics feature and patient data-based model, the single-image-based model, and the multi-patch-based model. The radiomics features are combined with the patient data to be used as input data for the random forest, whereas the EBUS images are used as input data to the other two CNN models. The performance of the proposed framework was evaluated on a set of 200 EBUS images consisting of 124 malignant lesions and 76 benign lesions. The experimental results show that the accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve are 95.00%, 100%, 86.67%, 92.59%, 100%, and 93.33%, respectively. This framework can significantly improve the pulmonary lesion classification.

This paper proposes an algorithm to extract food and beverage information from Thai receipt images using deep learning model and the proposed template-based information extraction. This information includes item name, quantity, price, and other details which are the sub-total, VAT, non-VAT, service charge and discount. The proposed algorithm consists of four steps: preprocessing, menu item template detection, optical character recognition, and information extraction. The experimental results show that the proposed algorithm using YOLOv8 for menu item template detection achieved mAP of 0.835, whereas OCR and the proposed template-based information extraction achieved 87.37% accuracy on item price extraction and 70.45% accuracy on VAT and other payment details.

Water is the key factor for sustainable human life. In addition to having adequate water source, the quality of water is also important. Water that is safe for human must meet standard water quality; otherwise, it is useless even though there is plenty of water. Thus, water quality must be measured regarding its physical, chemical, and biological properties. The purpose of this study is to apply time series analysis to model and predict Biochemical Oxygen Demand (BOD) for water quality at four monitoring stations along Chaophraya River of Thailand. In this paper, we propose an α-trimmed ARIMA model which can be used to predict BOD value of the up-coming year using a collection of BOD data from the past. The main advantage of our proposed model is that it can be used with both seasonal and non-seasonal time series data. The model was evaluated on a set of BOD data that were collected during 1996–2013. The predicted BOD results are compared to the BOD results obtained from other three existing models and the results reveal that the relative errors of our proposed model are less than half of the relative errors of those three existing models.