Active learning is a technique for maximizing performance of machine learning with minimal labeling effort and letting the machine automatically and adaptively select the most informative data for labeling. Since the labels on records may contain sensitive information, privacy-preserving mechanisms should be integrated into active learning. We propose a privacy-preservation scheme for active learning using homomorphic encryption-based federated learning. Federated learning provides distributed computation from multiple clients, and homomorphic encryption enhances the privacy preservation of user data with a strong security level. The experimental result shows that the proposed homomorphic encryption-based federated learning scheme can preserve privacy in active learning while maintaining model accuracy. Furthermore, we also provide a Deep Leakage Gradient comparison. The proposed scheme has no gradient leakage compared to the related schemes that have more than 74% gradient leakage.

The inspection of packaging defects is a crucial aspect of maintaining the quality of industrial production, especially in the case of boxed products. This study introduces a novel approach for detecting physical defects in product packaging boxes by integrating image processing with deep learning, specifically transfer learning with two images as an input. The proposed method utilizes both top view and side view images of the packaging to determine its condition, a significant departure from the conventional single image input. Our approach incorporates 16 pre-trained model variants from EfficientNetV2, MobileNetV3, and ResNetV2 for transfer learning as feature extractors. The experimental findings demonstrate that the best model that leverages EfficientNetV2 variant achieves 100% accuracy and F1 score in terms of classification performance. However, the most optimal model in terms of classification performance and inference speed was the one that leveraged ResNetV2 variant. This model scored 95% accuracy and 95.24% F1 score, with an inference speed of 91 ms per prediction.

This paper presents an efficient and simple technique on reducing the false contour problem which often occurs in the JPEG decoded image. The false contour appears on JPEG decoded image while applying small quality factor. This problem induces unpleasant visual appearance. The proposed scheme exploits the usability of Halftoning-Based Block Truncation Coding (HBTC) approach on generating visual illumination to reduce the aforementioned problem. Three HBTC techniques, namely Ordered Dither Block Truncation Coding (ODBTC), Error Diffusion Block Truncation Coding (EDBTC) and Dot Diffused Block Truncation Coding (DDBTC), modify the DC components of all Discrete Cosine Transform (DCT) processed image on JPEG encoding stage. Experimental results show the proposed method superiority on JPEG false contour reduction. To further improve the JPEG decoded image quality, the proposed method utilizes the Gaussian kernel on replacing the DDBTC diffused kernel on spreading the error term. It assumes that the adjacent image blocks have a strong correlation in which high diffused coefficient should be applied on nearest adjacent neighbor. This paper extends the HBTC usability on JPEG false contour suppression.

Nurse’s scheduling in hospitals becomes a complex problem, and it takes time in its making process. There are a lot of limitation and rules that have to be considered in the making process of nurse’s schedule making, so it can fulfill the need of nurse’s preference that can increase the quality of the service. The existence variety of different factors that are causing the nurse scheduling problem is so vast and different in every case. The study is aimed to develop a system used as an equipment to arrange nurse’s schedule. The working schedule obtained will be checked based on the constraints that have been required. Value check of the constraint falsification used Simulated Annealing (SA) combined with cooling method of Probabilistic Cooling Scheme (PCS). Transitional rules used cost matrix that is employed to produce a new and more efficient state. The obtained  results showed that PCS cooling methods combined with the transition rules of the cost matrix generating objective function value of  new solutions better and faster in processing time than the cooling method exponential and logarithmic. Work schedule generated by the application also has a better quality than the schedules created manually by the head of the room.

Due to the widespread use of distributed data mining techniques in a variety of areas, the issue of protecting the privacy of sensitive data has received increasing attention in recent years. Privacy-preserving distributed data mining (PPDDM) focuses on decentralized data analysis without the disclosure of sensitive information from data owner. However, the previous PPDDM mostly works on a limited amount of labeled data. In contrast to the real world, unlabeled data is abundance and labeled data is scarce. The objectives of this paper are to study and to analyze privacy-preserving properties of semi-supervised learning (SSL) algorithm with the combination of labeled and unlabeled data, where data is distributed among multiple data owners. In this paper we propose a Privacy-preserving Distributed Data Mining (PPDDM) method by designing a reliable application of secure MPC to semi-supervised tri-training algorithms. We simulate the original tri-training algorithm and tri-training algorithm with secure MPC using a different types of classifiers and datasets. The simulation results show that tri-training in secure MPC has almost same accuracy compared to original tri-training algorithm. We also compare execution time in addition to performance evaluation of tri-training in secure and the original tri-training algorithms.

The Coronavirus Disease 2019 (COVID-19) pandemic has affected all aspects of healthcare in the world with infection rates increasing exponentially. The detection process that has been carried out so far through Reverse Transcription Polymerase Chain Reaction (RT-PCR) examination, has a low level of accuracy, examination time, and sensitivity. Detecting COVID-19 can also be accomplished through the utilization of radiographic images like chest X-rays (CXR) and Computed Tomography (CT) with a higher level of accuracy and a fast examination time. By implementing deep learning as a predictive method for detecting COVID-19, results can be obtained quickly with a better level of accuracy. However, in its development, not many studies have analyzed the impact of improving image quality, which has considerable influence on the accuracy of the image data tested. Therefore, this research was conducted to analyze the results of testing the transfer learning model of COVID-19 images and normal images using the images quality has improved through the image-enhancement process. The image enhancements used in the research are Contrast Limited Adaptive Histogram Equalization (CLAHE) and Unsharp Masking (UM). Then, the improved quality images are trained with pre-trained Convolutional Neural Networks (CNN) methods, ResNet34 and EfficientNet-B4. By using the images that have been improved using CLAHE, the accuracy which obtained in the ResNet34 model are 77%, and 90% with the EfficientNet model, respectively. Furthermore, the results on testing images that have been enhanced using the UM method obtained 81% accuracy on the ResNet34 model, and 96% on the EfficientNet-B4 model, respectively.