Data in the health sector are often lacking and unbalanced. It is because collecting data takes time and many resources. One example is sleep apnea data which takes about 8–10 hours to get data and uses specialized hardware like polysomnography (PSG). This study proposes a data augmentation technique to handle unbalanced data using DCGAN and several deep learning models such as 1D-CNN, ANN, LSTM, and 1D-CNN+LSTM as a classifier for apnea detection. The DCGAN architecture used is CNN on the generator and discriminator. DCGAN will create new synthetic data by mimicking the original dataset. This experiment uses a dataset from PhysioNet, the Apnea-ECG, and the MIT-BIH PSG Database. Furthermore, the dataset is preprocessed to remove noise, and the features are extracted manually. The test scenario is to create 10% synthetic data and 50% sleep apnea data to be added to the original dataset. Then compare the performance of multiple deep learning models before and after adding data. The results indicate that augmentation with DCGAN can improve the performance of almost all models, with the highest increase of 1.78% on the 1D-CNN+LSTM model and 4.80% on the LSTM model for the Apnea-ECG and MIT-BIH datasets, respectively.

Facial Expression Recognition (FER) systems are helpful in a wide range of industries, including healthcare, social marketing, targeted advertising, the music industry, school counseling systems, and detection in the police sector. In this research, using Deep Convolutional Neural Networks (DCNN) architecture, specifically from the EfficientNet family (EfficientNet-B0, Efficient-Net-B01, EfficientNet-B02, EfficientNet-B03, EfficientNet-B04, EfficientNet-B05, EfficientNet-B06, and EfficientNet-B07) has previously gone through a combined scaling process of combined dept, width and resolution. First, the previously frozen sublayer EfficientNet model was used for feature extraction. Next, the layer closer to the output layer is melted by several layers to be retrained in order to recognize the pattern of the CK+ and JAFFE data sets. This process is called the transfer learning technique. This technique is very powerful for working on relatively small data sets, namely CK+ and JAFFE. The main of this research is to improve the accuracy and performance of facial expression recognition models with a transfer learning approach using EfficientNet pre-trained with fine-tuning strategy. Our proposed method, specifically using the EfficientNet-B0 architecture, achieves superior performance for each of the CK+ and JAFFE datasets, achieving 99.57% and 100% accuracy in the test set respectively.