Methodology manuscript¶
The current two-column manuscript specifies the training-free spherical CAM object-hypothesis methodology and its pre-declared evaluation programme. It is a methods paper: it intentionally contains no experimental results or performance claim.
Read the executable From spherical CAMs to spatial object hypotheses tutorial and Semantic object association and localization API contract.
The source contains original TikZ diagrams for the problem definition, typed data flow, spherical region construction, and cross-view reduction. Dataset images, checkpoints, private panoramas, and generated experiment panels are not part of the manuscript directory.