Skip to content

Dataset · vapaau

Fish segmentation and measurement on a conveyor

vapaau/autofish

1,500 photos of fish on a conveyor belt with species annotations, masks, and manual length measurements.

What's inside

Inside are 1,500 high-quality images capturing 454 unique fish specimens on a conveyor belt. For each fish, a class label, ID, and manual length measurements are indicated, and the set contains a total of 18,160 instance segmentation masks. The data is divided into 25 non-overlapping groups of 14–24 fish each, simplifying the creation of training splits.

What you can do with it

  • Training instance segmentation models (such as Mask R-CNN or YOLO) for object counting and sorting in manufacturing
  • Creating computer vision systems for estimating fish sizes and identifying species
Tasks
Image segmentation, Mask generation
Languages
English
Size
up to 1K rows
Downloads this month
23K
Likes
4

How to load it

from datasets import load_dataset

ds = load_dataset("vapaau/autofish")