Research

Papers

PaperVenueDeadlineStatusSlack
First paper · height sampling
Height-aware point sampling for forest semantic segmentation
Undecided (IEEE GRSL letter or CVPR EarthVision workshop)—Planning—
P0 · Q1 review
3D perception of wildfire fuel structure: a systematic review
Ecological Informatics (first choice)15 Jan 2027Planning69 days
P1 · ICCV 2027
Occlusion-aware point selection and fusion for understory segmentation
ICCV 20278 Mar 2027 (est.)Planning87 days
P2 · NeurIPS 2027
Budgeted agent-guided refinement of 3D fuel structure, or a hide-and-check benchmark
NeurIPS 2027 (main track or Datasets and Benchmarks; decide by 15 Apr 2027)21 May 2027 (est.)Planning—
P3 · Q1 journal
Occlusion-aware fuel-structure estimation validated against ground-based laser scans
ISPRS Journal of Photogrammetry and Remote Sensing, or Remote Sensing of Environment15 Sep 2027Planning—
Thesis
MSc thesis
—30 Sep 2027 (est.)Planning—

Experiments

ExperimentStatusVerdict
E1: occlusion-aware point selectionPlanned—
E2: PointNet++ with a learned fusion modulePlanned—
E3: budgeted agent (restore, refine, stop)Planned—
HELIOS++ hide-and-checkRunning—
Notebook 01: PointNet and PointNet++ baselines on ForestSemantic-MSDone—
Notebook 02: sampling retention gateDoneNO-GO on ForestSemantic-MS