Codesota · Computer Vision · Few-Shot Image Classification · COCO 2017 StuffTasks/Computer Vision/Few-Shot Image Classification
Few-Shot Image Classification · benchmark dataset · EN

COCO-Stuff (COCO 2017 Stuff / COCO-Stuff 164K).

COCO-Stuff (COCO 2017 Stuff) augments the MS COCO dataset with dense pixel-wise annotations for "stuff" classes (amorphous background regions like sky, grass, road). The COCO-Stuff v2 release annotates all ~164K images in the COCO 2017 collection with 91 stuff classes (in addition to the 80 COCO thing classes), enabling large-scale semantic segmentation and scene-understanding research focused on stuff/thing interactions and context. The annotations were produced with an efficient superpixel-based protocol that leverages COCO thing masks. (Original COCO dataset: arXiv:1405.0312; COCO-Stuff paper/announcement: arXiv:1612.03716 / CVPR 2018.)

Paper Submit a result
§ 01 · Leaderboard

Best published scores.

No results indexed yet — be the first to submit a score.

No benchmark results indexed yet
§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result Read submission guide
What a submission needs
  • 01A public checkpoint or API endpoint
  • 02A reproduction script with frozen commit + seed
  • 03Declared evaluation environment (Python, deps)
  • 04One row per metric declared by this dataset
  • 05A contact so we can follow up on discrepancies