Modelpublic

imageomics/butterfly_segmentation_yolo_v8

sourceHugging Facemitupdated 10mo agoView on Hugging Face

1likes

Model Card

Model Card for butterflysegmentationyolo_v8

This model takes in an image of a butterfly (with or without body attached to wings) and segments out any existing hindwings and forewings, in addition to pictured equipment described below.

Model Details

yolov8mshear10.0scale0.5translate0.1fliplr0.0_best.pt is the butterfly segmentation model.

The butterfly segmentation model was trained on a dataset of 800 total images from the Jiggins, OM_STRI, and Monteiro datasets. The model architecture is based on YOLO v8 (yolov8m-seg.pt), which we fine-tune further on our dataset of 800 images.

Model Description

The model is responsible for taking an input image (RGB) and generating segmentation masks for all classes below that are found in the image. Data augmentations applied during training include shear (10.0), scale (0.5), and translate (0.1). The model was trained for 50 epochs with an image size of 256. Note that despite defining an image size of 256, the normalized masks predicted by yolo can be rescaled to the original image size.

Segmentation Classes

[pixel class] corresponding category

—[0] background
—[1] right_forewing
—[2] left_forewing
—[3] right_hindwing
—[4] left_hindwing
—[5] ruler
—[6] white_balance
—[7] label
—[8] color_card
—[9] body

Details

model.train(data=YAML, imgsz=256, epochs=50, batch=16, device=DEVICE, optimizer='auto', verbose=True, val=True, shear=10.0, scale=0.5, translate=0.1, fliplr = 0.0 )

Metrics

Class Images Instances mAP50-95 all 64 358 background 64 3 0.20946 rightforewing 64 58 0.9845 leftforewing 64 51 0.9682 righthindwing 64 59 0.95296 lefthindwing 64 50 0.93961 ruler 64 31 0.73608 whitebalance 64 18 0.90686 label 64 50 0.80865 colorcard 64 24 0.92653 body 64 14 0.78283

Developed by: Michelle Ramirez

How to Get Started with the Model

To view applications of how to load in the model file and predict masks on images, please refer to this github repository

Citation

BibTeX:

@software{Ramirez_Lepidoptera_Wing_Segmentation_2024,
  author = {Ramirez, Michelle},
  doi = {10.5281/zenodo.10869579},
  month = mar,
  title = {{Lepidoptera Wing Segmentation}},
  url = {https://github.com/Imageomics/wing-segmentation},
  version = {1.0.0},
  year = {2024}
}

APA:

Ramirez, M. (2024). Lepidoptera Wing Segmentation (Version 1.0.0) [Computer software]. https://doi.org/10.5281/zenodo.10869579

Acknowledgements

The Imageomics Institute is funded by the US National Science Foundation's Harnessing the Data Revolution (HDR) program under Award #2118240 (Imageomics: A New Frontier of Biological Information Powered by Knowledge-Guided Machine Learning). Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.