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

Microsoft COCO (Common Objects in Context) — 2017 Panoptic Segmentation.

Microsoft COCO 2017 Panoptic Segmentation is the COCO 2017 subset annotated for panoptic segmentation, a task that unifies instance segmentation for "thing" classes and semantic segmentation for "stuff" classes into a single per-pixel labeling. The 2017 release contains the standard COCO image splits (train2017 with ~118,000 images and val2017 with 5,000 images, ~123k images total) and panoptic annotations (JSON panoptic annotation files plus per-image panoptic PNG segment maps). Panoptic annotations provide both instance ids and semantic class ids so models can be evaluated on a single panoptic quality metric that accounts for both things and stuff. The dataset is derived from the Microsoft COCO collection introduced in Lin et al., "Microsoft COCO: Common Objects in Context" (arXiv:1405.0312 / ECCV 2014). Common distribution points and tools include the official COCO website (cocodataset.org), the cocodataset panoptic API (cocodataset/panopticapi on GitHub), and multiple Hugging Face dataset mirrors that expose the train/val splits and panoptic annotations (e.g., AISNP/COCO2017-panoptic).

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