Muhammad Umer Ramzan

Learning Structurally Consistent Representations for Multi-View Radar Semantic Segmentation

Ali Zia, Muhammad Umer Ramzan, Abdelwahed Khamis, Usman Ali et autres

Radar sensors provide reliable perception under adverse weather and lighting conditions, but their sparse, noisy, and weakly semantic measurements make dense semantic segmentation challenging. Most existing radar segmentation methods rely on grid-based encodings and pairwise interactions, which struggle to capture the higher-order …

au, pk (code pays fourni par la source)

HyperEcho: Structured Higher Order Representation for Automotive Radar Semantic Segmentation

Abdelwahed Khamis, Muhammad Umer Ramzan, Usman Ali, Ali Zia

au, pk (code pays fourni par la source)

Component-Aware Sketch-to-Image Generation Using Self-Attention Encoding and Coordinate-Preserving Fusion

Ali Zia, Muhammad Umer Ramzan, Usman Ali, Muhammad Ashraf Faheem et autres

Translating freehand sketches into photorealistic images remains a fundamental challenge in image synthesis, particularly due to the abstract, sparse, and stylistically diverse nature of sketches. Existing approaches, including GAN-based and diffusion-based models, often struggle to reconstruct fine-grained details, maintain spatial alignment, or …

Component-Aware Sketch-to-Image Generation Using Self-Attention Encoding and Coordinate-Preserving Fusion

Ali Zia, Muhammad Umer Ramzan, Usman Ali, Muhammad Ashraf Faheem et autres

Translating freehand sketches into photorealistic images remains a fundamental challenge in image synthesis, particularly due to the abstract, sparse, and stylistically diverse nature of sketches. Existing approaches, including GAN-based and diffusion-based models, often struggle to reconstruct fine-grained details, maintain spatial alignment, or …

Geometry-Aware Semantic Reasoning for Training Free Video Anomaly Detection

Ali Zia, Usman Ali, Muhammad Umer Ramzan, Hamza Abid et autres

Training-free video anomaly detection (VAD) has recently emerged as a scalable alternative to supervised approaches, yet existing methods largely rely on static prompting and geometry-agnostic feature fusion. As a result, anomaly inference is often reduced to shallow similarity matching over Euclidean embeddings, …

Geometry-Aware Semantic Reasoning for Training Free Video Anomaly Detection

Ali Zia, Usman Ali, Muhammad Umer Ramzan, Hamza Abid et autres

Training-free video anomaly detection (VAD) has recently emerged as a scalable alternative to supervised approaches, yet existing methods largely rely on static prompting and geometry-agnostic feature fusion. As a result, anomaly inference is often reduced to shallow similarity matching over Euclidean embeddings, …

au, pk (code pays fourni par la source)

Learning Structurally Consistent Representations for Multi-view Radar Semantic Segmentation

Ali Zia, Muhammad Umer Ramzan, Abdelwahed Khamis, Usman Ali et autres

au, pk (code pays fourni par la source)

Split-Fuse-Transport: Annotation-Free Saliency via Dual Clustering and Optimal Transport Alignment

Muhammad Umer Ramzan, Ali Zia, Abdelwahed Khamis, Noman Ali et autres

Salient object detection (SOD) aims to segment visually prominent regions in images and serves as a foundational task for various computer vision applications. We posit that SOD can now reach near-supervised accuracy without a single pixel-level label, but only when reliable pseudo-masks …

pk, au (code pays fourni par la source)

Intelligent Predictive Maintenance in Urban Rail Systems: Computer Vision Approaches for Enhanced Wear Detection

Ameer Tamoor Khan, Ahmad Jahanzaib, Abdul Rehman Khan, Muhammad Umer Ramzan et autres

This study explores the integration of deep learning techniques for predictive maintenance within urban rail systems, emphasizing advanced automated rail head wear detection utilizing Mask Region-based Convolutional Neural Network (Mask R-CNN) and You Only Look Once (YOLO)v8 models. The increasing traffic and …

dk, hk, pk, fi (code pays fourni par la source)

Split-Fuse-Transport: Annotation-Free Saliency via Dual Clustering and Optimal Transport Alignment

Muhammad Umer Ramzan, Ali Zia, Abdelwahed Khamis, Usman Ali

Salient object detection (SOD) aims to segment visually prominent regions in images and serves as a foundational task for various computer vision applications. We posit that SOD can now reach near-supervised accuracy without a single pixel-level label, but only when reliable pseudo-masks …

Sketch to image synthesis: harnessing deep learning techniques for realistic visual transformation

Ahmad Gull Liaqat, Muhammad Umair Yousaf, Muhammad Umer Ramzan, Usman Ali et autres

Crime poses a significant threat to society, with rising rates in developing countries often attributed to unidentified criminals and various social factors. Traditional black-and-white sketches used by law enforcement agencies pose challenges as witnesses struggle to definitively identify suspects. In order to …

pk (code pays fourni par la source)

Gated-Attention Feature-Fusion Based Framework for Poverty Prediction

Muhammad Umer Ramzan, Wahab Khaddim, Muhammad Ehsan Rana, Fatima Mehmood

pk, my (code pays fourni par la source)