This research aims to apply an electronic nose (e-Nose) identifying TB through exhaled breath combined with machine learning methods. For the e-Nose system, participants breath normally into the e-Nose chamber. The e-Nose classified the exhaled breath into Tuberculosis Suspect (TBS) and Healthy Person (HP). The dataset was created by collecting directly after pulmonologist confirmed by using chest x-ray images. The machine learning methods were used to train and classify the exhaled breath. E-Nose models were created by using Support Vector Machine (SVM), Random Forest, and XGBoost. Based on experiments, the exhaled breath accuracy is 92 % for SVM, 88.24% for XGBoost, 94.87% for ANN, and 84.24% for the random forest.

In this paper, we describe an electronic nose (e-Nose) capable of classifying the aroma of alcoholic beverages. The novelty of this research is using signal processing for initial feature extraction from a sensor and then the use of deep learning to identify patterns of alcoholic beverage aromas. The sensor array was formed by nine types of metal oxide semiconductor sensors. The dataset was formed by images of standard deviations and correlation coefficients for processed signals from the e-Nose sensors. Thus, two patterns were generated. The first pattern came from a polar-chart image of the processed signals' standard deviations. The second pattern was produced by correlation coefficient converted into 3D heat-map images. The image size is 256 × 256 pixels. The convolutional architecture for fast feature embedding framework then trained GoogLeNet network using the dataset images. The training process was configured for 300 epochs and 0.0001 learning rate. The GoogLeNet network model from deep learning was compared with the AlexNet network model. The final classification was based on two patterns of prediction. The true label is used if the prediction accuracy value exceeds 70 %. The result is true only if both 3D heatmap and polar-chart prediction have true labels. The aroma detection accuracies of the GoogLeNet model are 85.0% for polar-chart and 85.416% accuracy for 3D heat-map. The aroma identification accuracy of AlexNet model are 85.0% for polarchart and 85.416% accuracy for 3D heat-map.

The non-contact healthcare system is a system that can avoid germ infection, and can also provide comfortable and convenient health care services for the caregivers. In the current thermal imaging research, local or small area images are used to represent the overall temperature information of the participants, but the temperature feature of the face should not only be used in a small part and ignore other parts. The facial thermal image can show clear temperature feature but is not conducive to meaningful feature extraction. Therefore, this research proposed a novel facial thermal image feature extraction method, which is used facial landmarks to detect and cut 12 blocks to establish a new feature matrix based on color mean values and standard deviation values. It can establish clear features on facial thermal images. The core part of the proposed healthcare system is the use of a deep learning framework, which is based on CAFFE under the DIGITS platform. The CAFFE runs the classic CNN, GoogLeNet. Based on the acquired images and new feature types, four models were trained, which were used for the raw RGB image, raw thermal image, RGB feature image, and thermal feature image. In the experiment, 800 images were used for training and validation, and 200 images were used for testing. An additional 40 images were used for random testing. The experimental results show that RGB images cannot be effectively used, thermal images can effectively predict the health status, and thermal feature images have the highest prediction accuracy.

Injection molding is one of the polymer molding methods. Product quality mainly can be affected by temperature and pressure. To observe the process of the melt forming in the mold, it is required to monitor the signal with the sensor. To achieve the purpose of robust molding, this research provides a new design-process window architecture based on machine learning of the injection product quality assessment. The process window is also known as the forming area map. The monitored value for defective products will be above or below the variable limit. The process is set to the center of this window so that any variation within the window will still result in an acceptable product. There are some potential problems in the current process. First, the conventional process window uses a small number of process variables as monitoring indicators, and it cannot provide stronger monitoring indicators. Second, most of the conventional process windows are 2-D and there are many monitoring windows, so the information that can be monitored is limited. Therefore, product quality cannot be effectively monitored without proper selection of important factors. Finally, the current process window can only provide information in the process and cannot make product quality prediction. In order to solve the above problems, a machine learning method is proposed to design the process window. The random forest classifier and regressor are used to predict the product quality. In order to improve the usability of the factor, the design of the proposed new variable is derived from the classical formula of thermodynamics (heat conduction). Using existing factors and proposing time-based feature changes can better represent the molding process as a whole. Based on the experimental results, the proposed variables can effectively increase the predictive performance of random forest model for quality assessment. The research results make a satisfactory result to the quality assessment of industrial injection molding. The accuracy of the prediction results is 100%. Moreover, it can be found from the 3-D process window that the data are very concentrated and can be clearly classified by the hyperplane, which is defined as the shortest distance between two clusters, and this space is the boundary of the process window. Therefore, compared with the past, the judgment boundary based on the random forest process window is no longer a simple boundary. Random forests can provide more accurate visualization boundaries.

Facial recognition based on visible-light RGB image is a mature technique. With the continuous progress of civilization, the protection against personal privacy and the effectiveness of the system is more stringent. In the past research, the distance from facial features was used as the de-identified biological feature, but this could not solve the problem of using a photo to deceive the system. Therefore, we used thermal image to convert into features by using the proposed feature extraction method, and uses deep learning, random forest, and ensemble learning to build a face prediction model. The proposed feature extraction method cuts the facial image (RGB and thermal image) into 12 and 48 blocks respectively as well as regenerates the feature image and the feature matrix. Based on the experimental results, accuracy of RGB image only with 0.834, feature image with 0.953, and feature matrix with 0.967. The feature images and feature matrices produced by the proposed feature extraction method can achieve better prediction performance. High accuracy of thermal imaging can solve the problem of fake photos and the issue of personal privacy de-identification could be solved.

People’s daily life is easily affected by mental stress, which can lead to mental illness in the long term. The current mental stress detection process is cumbersome, and the development of rapid assessment methods will make a great contribution to medical care. In view of this, this study used a pulse oximeter to obtain non-invasive photoplethysmography (PPG) signals, the measurement information was analyzed using heart rate variability (HRV), and the Poincaré plot of the heartbeat cycle was output. Poincaré maps are used as input to deep learning to perform conditional prediction of mental stress. Finally, compare the results of conventional HRV and deep learning. From the experimental results, the classification of the feature of the PPG signal (Poincaré plot) by the model is a meaningful and good result.