Skip to content

DZPeru/fish-datasets

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

32 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Fish Datasets: Segmentations & labelling

Datasets of fish for deep learning. A standarized way to create a public dataset of fish (images -> labelling -> dataset exports). If you want to add more images create an issue.

Models trained by the datasets (weights and demo available)

Datasets

Datasets Exports Standardized

Thanks to Roboflow, we can export for different formats:

  • Powered by Roboflow (All images resized to 416x416): JSON (COCO, CreateML), XML (Pascal VOC), TXT (YOLO Darknet, YOLO v3 Keras, YOLO v4 PyTorch, YOLO v5 PyTorch), CSV (Tensorflow Object Detection, RetinaNet Keras, Multiclass Classification), Others (Tensorflow TFRecord).

  • Powered by CVAT (All images original size): JSON (COCO, Datumaro), XML (CVAT Image, CVAT Video, LabelMe, PASCAL VOC), TXT (MOT, YOLO), JPG,PNG (ImageNet, Segmentation Mask), Others (Tensorflow TFRecord).

Choose just one zip file for one model, select the zip file accord to your model or download all the exports on release section

Date Images Classes Download dataset & Code (TR70/VL20/TS10)
2020-XX-XX 7500 (2500) +1 FISH9005 - Coming soon ...
2021-02-15 2400 (1000) 1 FISH9004 - less errors ✔️
2020-12-15 2400 (1000) 1 FISH9003 ✔️
2020-12-11 723 (301) 1 FISH9002 ✔️
2020-10-18 301 (301) 1 FISH9001 ✔️
  • By now just segmented salmon. But the list to complete is here.

Downloading and using dataset

Example code to use only with Roboflow notebooks to use the repo and locate correctly.

# Instead of doing:
!curl -L "ROBOFLOW LINK" > roboflow.zip; unzip roboflow.zip; rm roboflow.zip
# Use:
!gdown --id "ID OF THE GOOGLE DRIVE ZIP" > roboflow.zip; unzip roboflow.zip; rm roboflow.zip

Resources [Last Update 2020-12-23]

Keywords to find these datasets: fish, trout

List of resources:

List of datasets found

Contributing

There's a guide to help creating this dataset standardized and exported to a multiple standards. Click here

License

This repository is under MIT LICENSE.