Uzma Akram

Papers

Total Citations

H-Index

About

Uzma Akram is a computer vision researcher whose work focuses on advancing human-computer interaction through hand gesture and pose recognition. Her most cited paper, "Hand Pose Recognition Using Parallel Multi Stream CNN" (2021, 11 citations), introduces a novel deep learning architecture that processes multiple spatial and temporal streams simultaneously to improve the accuracy of hand pose estimation. This contribution addresses a critical challenge in enabling touchless, intuitive control for applications ranging from sign language interpretation and robot manipulation to smart surveillance and gaming. By designing a parallel multi-stream convolutional neural network, Akram’s work enhances the robustness of gesture recognition in real-world environments, reducing reliance on traditional input devices. Her research sits at the intersection of deep learning, human-computer interaction, and assistive technology, with potential implications for accessibility and automation. Though early in her citation impact, Akram’s innovative approach to multi-stream architectures marks a meaningful step toward more natural and responsive human-machine interfaces, positioning her as an emerging voice in applied computer vision.