Research Article
Enhanced Access Control Management System in an Edge Computing Environment Using Bimodal Biometrics Technique
Issue:
Volume 9, Issue 4, December 2026
Pages:
138-153
Received:
15 September 2026
Accepted:
24 September 2026
Published:
14 October 2026
Abstract: This study presents an enhanced access control management system for edge computing environments using bimodal biometrics, contextual risk analysis, and Chinese Wall policy enforcement. The work addresses a limitation in context-only access control frameworks, namely their inability to confirm that the person requesting access is the genuine enrolled user. To close this identity-verification gap, the proposed system integrates face recognition, fingerprint recognition, and contextual decision support into a unified edge-aware authentication workflow. The implementation was developed in Python using Flask, with MySQL for transactional data storage and IPFS for content-addressed storage of biometric-related artifacts. A hybrid dataset strategy was adopted. Public datasets such as VGGFace2 and SOCOFing supported preliminary experimentation, while locally captured biometric samples were emphasized for operational training and evaluation in order to reduce domain mismatch. A one-dimensional convolutional neural network was used for contextual access analysis, while transfer-learning-based convolutional neural networks were used for face and fingerprint recognition. The datasets were split into 70% training and 30% testing. Experimental results showed strong performance for the biometric and contextual modules. The face model achieved about 0.96 accuracy, the fingerprint model achieved about 93.36% classification accuracy with 95.83% precision, recall, and F1-score for the enrolled-user fingerprint class, and the contextual Conv1D model achieved 84.08% test accuracy. System-level testing further showed that integrating biometric evidence with contextual intelligence and policy enforcement improved access-decision reliability over the earlier context-only model. The study concludes that combining bimodal biometrics with edge-based contextual reasoning can provide stronger identity assurance, better real-time responsiveness, and more dependable access control for distributed computing environments.
Abstract: This study presents an enhanced access control management system for edge computing environments using bimodal biometrics, contextual risk analysis, and Chinese Wall policy enforcement. The work addresses a limitation in context-only access control frameworks, namely their inability to confirm that the person requesting access is the genuine enroll...
Show More
Research Article
Detecting AI-Generated Images via Fine-Tuned SigLIP 2:
A Sigmoid-Contrastive Vision-Language Encoder for Binary Authenticity Classification
Kanwerjit*,
Gaurav Deep
,
Sumandeep Kaur
Issue:
Volume 9, Issue 4, December 2026
Pages:
154-164
Received:
16 August 2026
Accepted:
14 September 2026
Published:
20 October 2026
DOI:
10.11648/j.ajcst.20260904.12
Downloads:
Views:
Abstract: The accelerating sophistication of generative artificial intelligence has produced widespread availability of photorealistic synthetic imagery, creating serious challenges for digital media integrity, forensic evidence reliability, and public trust in visual information. In response, developing robust yet computationally feasible automated detection systems has become a research priority. This study examines whether fine-tuning Sigmoid Loss for Language-Image Pre-training SigLIP2- a multilingual vision-language encoder- enables reliable discrimination between human-captured photographs and algorithmically generated images. The SigLIP2 base variant (google/siglip2-base-patch16-224) was adapted by replacing its contrastive projection head with a single linear classification layer applied to the pre-trained vision encoder. Evaluation on a held-out, balanced test set of 20,000 images yielded a classification accuracy of 98.90%, with macro-averaged and weighted F1-scores of 0.9890 each. These results demonstrate that the representational capabilities acquired by SigLIP2 through its staged pretraining- including sigmoid-based contrastive alignment, decoder-enforced spatial localization, and late-stage self-distillation- transfer effectively to the forensic discrimination of authentic versus synthetic imagery. The proposed framework is fully reproducible, and its potential applications include automated content moderation, digital forensics for legal contexts, investigative journalism verification workflows, and social media integrity systems.
Abstract: The accelerating sophistication of generative artificial intelligence has produced widespread availability of photorealistic synthetic imagery, creating serious challenges for digital media integrity, forensic evidence reliability, and public trust in visual information. In response, developing robust yet computationally feasible automated detectio...
Show More
