New academic paper in IEEE

The paper Face-looking Image Recognition, presented at the 27th Signal Processing and Communications Applications Conference, investigates recognition performance under constrained yet naturally occurring head pose and gaze scenarios (“face‑looking” conditions). By narrowing variation while retaining ecological validity, such scenarios can illuminate algorithmic sensitivity to subtle alignment and illumination shifts. The work adds incremental evidence toward balancing data diversity with controlled acquisition in model evaluation pipelines. Access the publication here.