Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68778
Volume:
10 |
11 |
Nov 2026
Pages:
1-7
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.

Computer Science and Engineering (AI & ML)
KLS Vishwanathrao Deshpande Institute of Technology, Haliyal
Computer Science and Engineering (AI & ML)
KLS Vishwanathrao Deshpande Institute of Technology, Haliyal
Computer Science and Engineering (AI & ML)
KLS Vishwanathrao Deshpande Institute of Technology, Haliyal
Computer Science and Engineering (AI & ML)
KLS Vishwanathrao Deshpande Institute of Technology, Haliyal
Computer Science and Engineering (AI & ML)
KLS Vishwanathrao Deshpande Institute of Technology, Haliyal
Road accident investigation requires systematic examination, organisation and documentation of photographic evidence collected from accident scenes. Traditional methods depend largely on manual inspection, which can be time-consuming when multiple photographs are involved. This paper presents AEGIS AI, an Android-based road accident evidence analysis system developed to support investigators through AI-assisted image analysis and structured evidence management. The application is developed using Flutter, Dart and Kotlin, with SQLite for local data storage and TensorFlow Lite for on-device inference. An experimental MobileNetV3 Small classifier is integrated for accident-related image classification, while EfficientDet-Lite0 identifies supported objects in photographs. The system incorporates accident case management, photographic evidence acquisition, image-quality assessment and SHA-256 hashing for evidence integrity verification. EvidenceLens enables detailed image inspection, while SceneCompare supports comparison of photographs captured from different viewpoints. Human verification and PDF report generation are integrated to support the review and documentation of findings. Functional testing on a Realme 11 smartphone running Android 15 confirmed the operation of the implemented application modules, including image navigation, object detection overlays and evidence inspection. The system distinguishes between accident evidence detected, no accident evidence detected and inconclusive outcomes. Although the integrated workflow was functionally tested, quantitative model accuracy and reliability require further evaluation using representative accident datasets. AEGIS AI demonstrates an approach to combining mobile computer vision with digital evidence management while retaining human responsibility for final interpretation.
KEYWORDS
AEGIS AI: An AI-Based Road Accident Evidence Analysis System was developed to support the systematic examination, management and documentation of photographic accident evidence. The Android application integrates accident case management, evidence acquisition, image-quality assessment, AI-assisted classification, object detection, evidence inspection, multi-image comparison, integrity verification, human review and PDF report generation within a unified workflow. The system uses MobileNetV3 Small and EfficientDet-Lite0 to assist in identifying accident-related visual information and detecting supported objects in photographs. EvidenceLens and SceneCompare provide additional facilities for examining individual images and comparing multiple photographs. SHA-256 hashing supports evidence integrity verification, while human verification ensures that AI-generated observations can be reviewed before documentation. Functional testing on an Android smartphone demonstrated the operation of the implemented application modules. However, further evaluation using representative accident datasets is necessary to establish quantitative model accuracy, reliability and performance under different environmental conditions. The system is intended to assist evidence examination rather than independently determine accident causes or legal responsibility. Overall, AEGIS AI demonstrates the practical integration of mobile computer vision and digital evidence management. By combining automated analysis with human oversight, the application provides a structured approach to photographic accident evidence examination and establishes a foundation for further research and improvement.
[1] S. Porwal and P. Singh, “Identification of collided vehicles in Indian traffic accidents using hierarchical deep learning framework,” in Proc. 21st Int. Conf. Computer Vision Theory and Applications (VISAPP), vol. 1, 2026, pp. 389–396, doi: 10.5220/0014290000004084. [2] B. M. Gurusamy, P. R. Sanikommu, and G. Muthurasu, “Real-time road accident detection and severity assessment using IoU and deep learning models,” Journal of Computational and Cognitive Engineering, 2025, doi: 10.47852/bonviewJCCE52025908. [3] K. Pawar and V. Attar, “Deep learning based detection and localization of road accidents from traffic surveillance videos,” ICT Express, vol. 8, no. 3, pp. 379–387, 2022, doi: 10.1016/j.icte.2021.11.004. [4] B. Guttman, D. R. White, S. Williams, and T. Walraven, “Digital evidence preservation: Considerations for evidence handlers,” National Institute of Standards and Technology, NIST IR 8387, 2022, doi: 10.6028/NIST.IR.8387. [5] National Institute of Standards and Technology, “Secure Hash Standard (SHS),” Federal Information Processing Standards Publication 180-4, 2015, doi: 10.6028/NIST.FIPS.180-4. [6] L. Verdoliva, “Media forensics and DeepFakes: An overview,” IEEE Journal of Selected Topics in Signal Processing, vol. 14, no. 5, pp. 910–932, 2020, doi: 10.1109/JSTSP.2020.3002101. [7] A. Howard et al., “Searching for MobileNetV3,” in Proc. IEEE/CVF Int. Conf. Computer Vision (ICCV), 2019, pp. 1314–1324.
Publication Date
11/6/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
4
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...