Article Per Year (5 Year)
p-Index From 2021 - 2026
0.61
P-Index
This Author published in this journals
All Journal RABIT: Jurnal Teknologi dan Sistem Informasi Univrab JOURNAL OF APPLIED INFORMATICS AND COMPUTING
Wildanil Ghozi
Unknown Affiliation
Computer Science & IT Engineering

Published : 3 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 3 Documents
Search

 1 
Performance Analysis of Deep Learning Model Quantization on NPU for Real-Time Automatic License Plate Recognition Implementation Alexander, Daniel; Wildanil Ghozi
Journal of Applied Informatics and Computing Vol. 9 No. 4 (2025): August 2025
Publisher : Politeknik Negeri Batam

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30871/jaic.v9i4.9700

Abstract

Neural Processing Units (NPUs) are dedicated accelerators designed to perform efficient deep learning inference on edge devices with limited computational and power resources. In real-time applications such as automated parking systems, accurate and low-latency license plate recognition is critical. This study evaluates the effectiveness of quantization techniques, specifically Post-Training Quantization (PTQ) and Quantization-Aware Training (QAT), in improving the performance of YOLOv8-based license plate detection models deployed on an Intel NPU integrated within the Core Ultra 7 155H processor. Three model configurations are compared: a full-precision float32 model, a PTQ model, and a QAT model. All models are converted to OpenVINO’s Intermediate Representation (IR) and benchmarked using the benchmark_app tool. Results show that PTQ and QAT significantly enhance inference efficiency. QAT achieves up to 39.9% improvement in throughput and 28.6% reduction in latency compared to the non-quantized model, while maintaining higher detection accuracy. Both quantized models also reduce model size by nearly 50 percent. Although PTQ is simpler to implement, QAT offers a better balance between accuracy and speed, making it more suitable for deployment in edge scenarios with real-time constraints. These findings highlight QAT as an optimal strategy for efficient and accurate license plate recognition on NPU-based edge platforms.
OPTIMASI KINERJA SISTEM DETEKSI INTRUSI MENGGUNAKAN HYBRID XGBOOST DAN ARSITEKTUR DEEP LEARNING EFISIEN Bernandiko Priyambodo; Wildanil Ghozi; Fauzi Adi Rafrastara
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 11 No 1 (2026): Januari
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v11i1.7182

Abstract

Rapid digital transformation has expanded the cyber attack surface, demanding a responsive and reliable Intrusion Detection System (IDS). The main obstacle in developing Deep Learning-based IDS is the high dimensionality of network traffic features and data class imbalance, which can trigger excessive computational loads. This study aims to develop an efficient and robust IDS model for detecting various types of cyber attacks using the CIC-IDS2017 dataset. The proposed method applies a hybrid approach integrating the XGBoost algorithm for high-importance feature selection to reduce data dimensionality, alongside a Multilayer Perceptron (MLP) architecture for classification. This research explores various neural network depth configurations combined with tanh and SELU activation functions to handle data non-linearity. Model performance is evaluated based on standard classification metrics as well as operational security metrics. Experimental results demonstrate that the proposed model achieved an accuracy of 99.39% with a low False Alarm Ratio (FAR) and Attack Miss Ratio (AMR) of 2.90%. This study contributes by presenting an intrusion detection framework capable of balancing architectural complexity and computational efficiency for implementation in modern network environments.
DETEKSI DAN MITIGASI SERANGAN POST-EXPLOITATION PADA LINGKUNGAN CONTAINER LINUX MENGGUNAKAN CROWDSEC DAN AUDITD Julang Tahta Pratangga; Chaerul Umam; L. Budi Handoko; Wildanil Ghozi
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 11 No 2 (2026): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v11i2.8461

Abstract

Containerized environments present significant security challenges due to their shared kernel architecture, which may be exploited as an entry point for infiltration attacks. One common vulnerability is Command Injection, enabling post-exploitation activities that are difficult to detect using conventional signature-based mechanisms. This study aims to implement and evaluate an active mitigation mechanism based on behavioral analysis by integrating the Linux Audit Daemon (Auditd) and CrowdSec within a Podman container environment. The research adopts an Experimental Security Testing approach by developing a Custom Process Bouncer that specifically monitors the execve system call to identify process relationships and automatically terminate malicious processes. Test results against four detection scenarios demonstrate that the proposed mechanism successfully reconstructed the attack process chain through PID cascade analysis and detected six malicious processes generated during the Command Injection scenario. All identified processes were automatically terminated with precision, causing the reverse shell session to be interrupted. These findings conclude that the integration of Auditd, CrowdSec, and the Custom Process Bouncer effectively neutralizes post-exploitation activities reliant on binary file execution, though further development is required to address fileless execution tactics.
Co-Authors Alexander, Daniel Bernandiko Priyambodo Chaerul Umam Fauzi Adi Rafrastara Julang Tahta Pratangga L. Budi Handoko