New Paper: “Object-Level Detection of Hand-Drawn Annotations in Participatory Sketch Maps Using Paired Clean and Annotated Basemaps”

A new paper presented at the AGILE 2026 conference investigates how hand-drawn annotations in participatory sketch maps can be automatically extracted and digitized. The authors formulate annotation extraction as an object-level detection task using YOLOv9e-based detectors, and use paired clean and annotated basemaps in the Siamese dual-image change detection setting. Their results suggest that this method can support the digitization of spatial information created by communities.