Coming soon: This challenge focuses on scene understanding from low-resolution, wide instantaneous field-of-view (IFoV) SPAD sensors (ams TMF8828, ST VL53L4CX) operating in both line-of-sight (LoS) and non-line-of-sight (NLoS) regimes. Participants will address three key tasks: (1) single-view 3D reconstruction from very low-resolution LoS transients (e.g., 3×3 or 8×8 SPAD arrays), where each pixel integrates light over a broad spatial region; (2) performing object localization and classification using multi-bounce NLoS signals; and (3) generating realistic simulated NLoS data of low-cost LiDARs, including accurate modeling of transient responses and sensor noise. The goal is to push the limits of active sensing in constrained, low-cost hardware settings by combining inverse problems and learning-driven approaches.