Naoufel Werghi

This work proposes an innovative solution that combines a lightweight vision-language encoder with a compact multi-layer perceptron to predict combustion efficiency directly from low-cost thermal video footage to achieve high combustion efficiency in flare stacks.

Organizing longitudinal chest X-ray evidence by anatomical regions dramatically boosts both language quality and reasoning accuracy in medical AI applications.

Falcon redefines threat assessment in X-ray screening by shifting from object detection to relational reasoning about component interactions.

Enforcing temporal validity during memory updates can stabilize visual tracking systems and lead to unprecedented performance gains across nine benchmarks.

Polycepta's recursive appearance state estimation leads to a remarkable 92.27% MOTA on the KITTI benchmark, outperforming traditional static methods.