Journal of Computer and Forensic Sciences

Comparative Analysis of Machine Learning Models for Real-Time Object Detection

2026-04-30T16:34:16+02:00

: Object detection is a fundamental task in computer vision with applications ranging from autonomous driving, industrial automation and medical imaging. This report presents a comparative analysis of six well-known object detection models, precisely three small models for edge computing and three large models likely more suited for usage on high-performance systems. The models YOLOv10-Nano, MobileNetV3-SSDLite, EfficientDet-D0, Faster R-CNN, YOLOv10-Large and DETR were evaluated and compared based on their performance in terms of inference speed, accuracy and computational efficiency. The evaluation is conducted through both literature-based benchmarks and empirical tests on two different systems: an Apple Silicon M1 Pro-based system and an NVIDIA RTX 3080Ti-powered computer. Results show that YOLOv10 models consistently outperform the other models in real-time object detection as well as achieving superior accuracy in general while maintaining significantly lower inference times. The analysis further highlights compatibility issues with certain hardware, particularly focusing on PyTorch's MPS backend on Apple Silicon, which leads to serious performance drops in some models. The findings highlight the importance of choosing the right model and appropriate hardware for specific application scenarios.

Heart_Rate_Estimation_Using_Wearable_Sensors_and_MachineLearning

2026-04-30T16:36:08+02:00

This research explores the development of a heart rate estimation system that integrates wearable sensors with machine learning techniques to achieve accuracy, low cost, and real-time performance. The project aims to build two-stage phases: a software pipeline for model training and a hardware framework for real-world testing. In the first phase, various machine learning algorithms are trained and fine-tuned using the publicly available PPG DaLiA dataset, which contains physiological data collected during everyday activities. The training process focuses on optimizing performance across different model architectures and configurations. The second phase involves implementing the trained model on a real-time embedded system. An ESP32 microcontroller serves as the central unit to collect data from multiple sensors, including electrocardiography (ECG), photoplethysmography (PPG), galvanic skin response (GSR), temperature, and a 3-axis accelerometer. This data is transmitted wirelessly for preprocessing and inference. The user will see their final predicted heart rate on both an OLED display and a user interface (UI) dashboard.

Application of Time Series Algorithms in Cybersecurity

2026-04-30T16:38:00+02:00

Abstract: This paper explores the application of time series algorithms to enhance anomaly detection in cybersecurity. Windows log files such as PowerShell Operational, Windows Defender, Firewall, System, and others were analyzed, focusing on those with the highest informational potential and data volume. Various models were used: exponential smoothing (Holt-Winters), Prophet, Fourier analysis, and Kalman filter for modeling seasonal, periodic, and linear patterns in system events. Advanced methods include LSTM and GRU neural networks, as well as ensemble algorithms like Random Forest and XGBoost, which demonstrated high accuracy in detecting unusual behavior. Special emphasis was placed on dynamic models such as Bayesian Structural Time Series to understand system states over time. Experiments show that applying multiple models enables a robust and adaptive approach to log analysis, especially for early detection of attacks and deviations from norms. The proposed framework highlights the importance of predictive analytics in preventive cybersecurity and provides a foundation for developing intelligent systems for real-time monitoring and response.