Khalifa University
The CVML Lab at Khalifa University advances robust computer vision and machine learning, developing intelligent visual systems for real-world applications in security and surveillance, biometrics, medical imaging, remote sensing, and autonomous systems.
Researching intelligent visual and audio analytics for security, environmental monitoring, and automated inspection across complex real-world environments. This includes X-ray/CT threat detection, crowd tracking, flare analysis, and robust speech-based identification in noisy public spaces.
| 📘 Title / Paper | 📝 Description | 🗂️ Repository | 📄 Paper | 🏅 Venue | 🎬 Demo |
|---|---|---|---|---|---|
| DERA: Detached Edge-Residual Adaptation for Prohibited Item Detection | An edge-aware residual adaptation framework for prohibited-item detection in X-ray security images. | DERA · Weights | Coming soon | - | Project Page |
| FALCON: Functional Assembly and Language for Compositional Reasoning in X-ray | A functional assembly and vision-language framework for compositional reasoning in X-ray security imagery. | FALCON · Weights · Benchmark | arXiv | ECCV 2026 | Project Page |
| A Benchmark Dataset for Concealed Improvised Explosive Device Detection in X-ray Security Imaging | Novel IED dataset with diverse IED types— homemade explosives, batteries, and modified devices such as laptops, mobile phones, pagers, and walkie-talkies. | IEDXray | Scientific Data (2026) | - | |
| STING-BEE: Towards Vision-Language Model for Real-World X-ray Baggage Security Inspection | A domain-aware vision-language model to support multimodal X-ray security tasks (Grounding, VQA, and scene understanding). | STING-BEE | -- | CVPR'25-Highlight | Demo |
| X-SSL: Self-supervised X-ray threat detection with zero-shot and multi-modal learning | A self-supervised X-ray threat localization framework. | X-SSL | Information Processing & Management (2026) | Demo | |
| Video anomaly detection in 10 years: A survey and outlook | Survey of deep learning video anomaly detection, including VLMs, challenges, methods, and future directions. | - | -- | Neural Computing and Applications | - |
| Training-Free VLM-Based Pseudo Label Generation for Video Anomaly Detection | Novel Training-free pseudo-label generation module for Weakly Supervised Video Anomaly Detection. | TFPLG | -- | IEEE Access | - |
--
We develop advanced computer vision and machine learning methods for the analysis of diverse medical imaging modalities, including MRI, CT, X-ray, fundus, and digital pathology. Our work aims to build robust computational frameworks that enhance diagnostic accuracy, support quantitative interpretation, and improve clinical decision-making across healthcare applications.
| 📘 Title / Paper | 📝 Description | 🗂️ Repository | 📄 Paper | 🏅 Venue | 🎬 Demo |
|---|---|---|---|---|---|
| DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation | A dynamic uncertainty-aware semi-supervised segmentation framework for robust learning under class imbalance. | DyCON | -- | CVPR'25 | https://youtube.com/dummy-demo |
| TriGAN-SiaMT: A Triple-Segmentor Adversarial Network with Bounding Box Priors for Semi-Supervised Brain Lesion Segmentation | Semi-supervised segmentation framework that integrates adversarial learning, multi-level consistency regularization, and bounding box priors. | TriGAN | -- | Pattern Recognition Letters (2025) | Demo |
| STING-BEE: Towards Vision-Language Model for Real-World X-ray Baggage Security Inspection | A domain-aware vision-language model to support multimodal X-ray security tasks (Grounding, VQA, and scene understanding). | Repo | -- | CVPR'25 | Demo |
Tools, utilities, sandbox code, and projects.
| 🗂️ Repository | 📝 Description | 🔗 Link |
|---|---|---|
| Vision-Based-Class-Attendance-System | Automated class attendance system based on facial recognition. | Repo |
| FaceNet | Real-time face recognition model. | Repo |
| Xray-to-Xray_Data_Augmentation | Synthesize X-ray baggage images. | Repo |
Interactive web-based demos developed by the CVML Lab, enabling real-time visualization and testing of our research systems.
| 🧠 Application | 📝 Description | 🔗 Demo |
|---|---|---|
| 🧳 Secure-X (X-ray Threat Detection) | Real-time X-ray baggage threat detection system for security screening and inspection. | Launch_Demo |
| 🧠 Brain Lesion Segmentation | Visualization and segmentation of brain lesions from medical imaging volumes (NIfTI format). | Launch_Demo |
Explore the research outputs and academic profiles of the CVML Lab:
For collaboration, research inquiries, or access to CVML Lab resources:
- 📧 Email: naoufel.werghi@ku.ac.ae
- 🌐 Website