To pursue a healthy lifestyle, people are increasingly concerned about their food ingredients. Recently, it has become a common practice to use an online recipe to select the ingredients that match an individual’s meal plan and healthy diet preference. The information from online recipes can be extracted and used to develop various food-related applications. Named entity recognition (NER) is often used to extract such information. However, the problem in building an NER system lies in the massive amount of data needed to train the classifier, especially on a specific domain, such as food. There are food NER datasets available, but they are still quite limited. Thus, we proposed an iterative self-training approach called semi-supervised multi-model prediction technique (SMPT) to construct a food ingredient NER dataset. SMPT is a deep ensemble learning model that employs the concept of self-training and uses multiple pre-trained language models in the iterative data labeling process, with a voting mechanism used as the final decision to determine the entity’s label. Utilizing the SMPT, we have created a new annotated dataset of ingredient entities obtained from the Allrecipes website named FINER. Finally, this study aims to use the FINER dataset as an alternative resource to support food computing research and development.

A respiratory disorder that attacks COVID-19 patients requires intensive supervision of medical practitioners during the isolation period. A non-contact monitoring device will be a suitable solution for reducing the spread risk of the virus while monitoring the COVID-19 patient. This study uses Frequency-Modulated Continuous Wave (FMCW) radar and Machine Learning (ML) to obtain respiratory information and analyze respiratory signals, respectively. Multiple subjects in a room can be detected simultaneously by calculating the Angle of Arrival (AoA) of the received signal and utilizing the Multiple Input Multiple Output (MIMO) of FMCW radar. Fast Fourier Transform (FFT) and some signal processing are implemented to obtain a breathing waveform. ML helps the system to analyze the respiratory signals automatically. This paper also compares the performance of several ML algorithms such as Multinomial Logistic Regression (MLR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), CatBoosting (CB) Classifier, Multilayer Perceptron (MLP), and three proposed stacked ensemble models, namely Stacked Ensemble Classifier (SEC), Boosting Tree-based Stacked Classifier (BTSC), and Neural Stacked Ensemble Model (NSEM) to obtain the best ML model. The results show that the NSEM algorithm achieves the best performance with 97.1% accuracy. In the real-time implementation, the system could simultaneously detect several objects with different breathing characteristics and classify the respiratory signals into five different classes.

Food recipe sharing sites are becoming increasingly popular among people who want to learn how to cook or plan their menu. Through online food recipes, individuals can select ingredients that suit their lifestyle and health condition. Information from online food recipes is useful in developing food-related systems such as recommendations and health care systems. However, the information from online recipes is often unstructured. One way of extracting such information into a well-structured format is the technique called named-entity recognition (NER), which is the process of identifying keywords and phrases in the text and classifying them into a set of predetermined categories, such as location, persons, time, and others. We present a food ingredient named-entity recognition model called RNE (recurrent network-based ensemble methods) to extract the entities from the online recipe. RNE is an ensemble-learning framework using recurrent network models such as RNN, GRU, and LSTM. These models are trained independently on the same dataset and combined to produce better predictions in extracting food entities such as ingredient names, products, units, quantities, and states for each ingredient in a recipe. The experimental findings demonstrate that the proposed model achieves predictions with an F1 score of 96.09% and outperforms all individual models by 0.2% to 0.5% in percentage points. This result indicates that RNE can extract information from food recipes better than a single model. In addition, this information extracted by RNE can be used to support various information systems related to food.

During the cooking process, the chemical structure of food ingredients undergoes changes, resulting in nonlinear variations in nutrient composition. This study develops a neural network model to predict nutrient changes in dishes caused by cooking processes. Built upon a feedforward network with a residual connection, the proposed model effectively learns the nonlinear characteristics of the changes by combining logarithmic scaling of the input and residual learning, which helps capture variations in the scale of the input. Experimental results demonstrate improved prediction accuracy for specific nutrients and confirme the model's ability to predict nutrient composition across diverse food ingredients. This research suggests potential applications in personalized diet planning and advancements in the food technology industry.

Prolonged viewing of videos in virtual reality environments induces visual and cognitive fatigue in users. This paper presents a method of predicting fatigue from watching virtual reality videos. The proposed method uses a camera to measure eye movements, pupil size changes, and blink rates and intervals, to estimation the degree of fatigue using a random forest. The prediction results are annotated using a standardized chart and then aligned to subjective fatigue evaluations for validation. This study will provide foundational data for enhancing VR user experience, by developing fatigue management capability and optimizing wearability and playability of VR contents

While recent advances in deep learning many existing approaches primarily emphasize textual descriptions and fail to fully capture the mathematical content of equations. To address this limitation, this study proposes a hybrid model that integrates a Long Short-Term Memory (LSTM) network with Bidirectional Encoder Representations from Transformers (BERT) for classifying equations into topic domains. In the proposed method, LaTeX-formatted equations are transformed into bigram token sequences for structural analysis by the LSTM encoder, while the surrounding textual context is processed by the BERT encoder to capture semantic meaning. The two encoded representations are concatenated and passed through a feed-forward neural network for final classification. Experiments on a high school mathematics dataset demonstrate that this approach outperforms standalone LSTM and BERT models, achieving 91.96% accuracy, 92.11% precision, 91.97% recall, and an F1-score of 0.92, confirming the effectiveness of jointly leveraging structural and contextual information for mathematical problem classification.

Asthma patients are prone to sudden respiratory failure, thus requiring a non-invasive and continuous monitoring method. This study investigates the use of Frequency-Modulated Continuous Wave (FMCW) radar to detect asthma by recording chest displacement signals without physical contact. The extracted respiratory features, including statistical descriptors and Mel-Frequency Cepstral Coefficients (MFCC), are analyzed using a machine learning classifier. To address class imbalance in the dataset, imbalanced learning strategies such as Synthetic Minority Oversampling Technique (SMOTE) and Adaptive Synthetic Sampling (ADASYN) are applied. Among the tested models, the LightGBM-ADASYN combination achieved the best performance with an F1 score of 92.95% on the training data and 92.94% on the testing data. These results demonstrate the effectiveness of integrating FMCW radar with imbalanced learning approaches for the early detection of asthma, enabling timely intervention and improving patient safety.

Age, gender, and race classification pose challenges across various domains, including computer vision, social sciences, and marketing. Despite advancements in deep and machine learning, convolutional neural networks (CNNs) remain crucial for addressing these tasks. This paper introduces an innovative approach utilizing CNNs with residual blocks to enhance accuracy and efficiency in age, gender, and race classification. Incorporating residual connections enables the model to capture both low-level and high-level features, improving classification accuracy while minimizing computational complexity. The residual blocks facilitate the learning of residual mappings, aiding gradient propagation and enabling successful training of deeper networks. Evaluation of the model on dataset FairFace demonstrates the model’s performance with accuracies of 56.3%, 94.6%, and 58.4% for age, gender, and race, respectively.

In this paper we propose a technique of correcting online handwriting recognition errors using morphemic lexicons and a set of confusion lists for mistaken characters collected from classification errors. Given a string of Hangeul characters returned by a baseline recognizer, we raise a set of alternative characters out of a confusion dictionary, one for each character in the returned string. Then we conduct a post-processing via a dynamic programming-based lattice search for a sequence of legal words while consulting morphemic lexicons. The proposed methods can handle short broken phrases in natural online handwriting of sentences. Experiment shows that, for handwriting strings of 3.09 characters long on average, the proposed method improved the phrase recognition accuracy from 55.43% before post-processing to 86.63%, an error reduction of 70.00%, with single hypotheses.

To pursue a healthy lifestyle, people are increasingly concerned about their food ingredients. Recently, it has become a common practice to use an online recipe to select the ingredients that match an individual’s meal plan and healthy diet preference. The information from online recipes can be extracted and used to develop various food-related applications. Named entity recognition (NER) is often used to extract such information. However, the problem in building an NER system lies in the massive amount of data needed to train the classifier, especially on a specific domain, such as food. There are food NER datasets available, but they are still quite limited. Thus, we proposed an iterative self-training approach called semi-supervised multi-model prediction technique (SMPT) to construct a food ingredient NER dataset. SMPT is a deep ensemble learning model that employs the concept of self-training and uses multiple pre-trained language models in the iterative data labeling process, with a voting mechanism used as the final decision to determine the entity’s label. Utilizing the SMPT, we have created a new annotated dataset of ingredient entities obtained from the Allrecipes website named FINER. Finally, this study aims to use the FINER dataset as an alternative resource to support food computing research and development.

A respiratory disorder that attacks COVID-19 patients requires intensive supervision of medical practitioners during the isolation period. A non-contact monitoring device will be a suitable solution for reducing the spread risk of the virus while monitoring the COVID-19 patient. This study uses Frequency-Modulated Continuous Wave (FMCW) radar and Machine Learning (ML) to obtain respiratory information and analyze respiratory signals, respectively. Multiple subjects in a room can be detected simultaneously by calculating the Angle of Arrival (AoA) of the received signal and utilizing the Multiple Input Multiple Output (MIMO) of FMCW radar. Fast Fourier Transform (FFT) and some signal processing are implemented to obtain a breathing waveform. ML helps the system to analyze the respiratory signals automatically. This paper also compares the performance of several ML algorithms such as Multinomial Logistic Regression (MLR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), CatBoosting (CB) Classifier, Multilayer Perceptron (MLP), and three proposed stacked ensemble models, namely Stacked Ensemble Classifier (SEC), Boosting Tree-based Stacked Classifier (BTSC), and Neural Stacked Ensemble Model (NSEM) to obtain the best ML model. The results show that the NSEM algorithm achieves the best performance with 97.1% accuracy. In the real-time implementation, the system could simultaneously detect several objects with different breathing characteristics and classify the respiratory signals into five different classes.