Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Jurnal Informatika Kaputama (JIK)
Rintan Nurhaliza
Unknown Affiliation
Computer Science & IT Control & Systems Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Transparansi Visi Komputer Sintesis Lanskap Dan Peta Jalan Explainable AI Rintan Nurhaliza; Muhamad Nur Ramdani
Jurnal Informatika Kaputama (JIK) Vol 10 No 2 (2026): Volume 10, Nomor 2, Juli 2026
Publisher : STMIK KAPUTAMA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59697/jik.v10i2.1387

Abstract

Although Deep Learning architectures in Computer Vision (CV) demonstrate exceptional classification performance, their black-box nature has triggered a crisis of user trust, particularly in the medical and cybersecurity sectors. Explainable Artificial Intelligence (XAI) emerges as a crucial paradigm for uncovering the rational logic behind these algorithmic decisions. This study aims to synthesize the current landscape and formulate a roadmap for XAI development in CV through a Systematic Literature Review (SLR). Using the PRISMA 2020 protocol, the literature search focused on 30 selected primary studies. The review results indicate that post-hoc agnostic methods (LIME, SHAP) and visual attribution based on saliency maps (Grad-CAM) dominate interpretation standards. There is also an architectural evolution toward hybrid models and Vision Transformers (ViT) to generate intrinsic explanations. XAI integration has been shown to improve generalization accuracy by eliminating feature noise and effectively countering manipulative attacks. However, high computational costs and the absence of human-centered evaluation metrics remain challenges. As a comprehensive solution to these issues, this study proposes a three-phase strategic roadmap to guide the development of future intelligent systems that are transparent, secure, and accountable, paving the way for more reliable advancements in visual technology.
Co-Authors Muhamad Nur Ramdani