PhenoLeaf-TS — Pretrained Models
Baseline checkpoints for the PhenoLeaf-TS benchmark (ECCV 2026), trained on the PhenoLeaf-TS dataset.
| File | Model | Framework | Task | Score |
|---|---|---|---|---|
IS-3_mask_rcnn_r50.pth |
Mask R-CNN R50-FPN | Detectron2 | Leaf instance segmentation | 73.2 mAP |
CL-6_swin_t.pth |
Swin-T | timm | Growth-stage classification | 91.7 % acc |
IS-1_yolov11_seg.pt |
YOLOv11-seg | Ultralytics | Segmentation + tracking detector | 68.2 mAP |
Usage
from phenoleaf_ts.models import load_model # see the GitHub repo
seg = load_model("IS-3", checkpoint="IS-3_mask_rcnn_r50.pth")
clf = load_model("CL-6", checkpoint="CL-6_swin_t.pth")
# or pull a file directly
from huggingface_hub import hf_hub_download
ckpt = hf_hub_download("basimazam/PhenoLeaf-TS-models", "IS-1_yolov11_seg.pt")
Citation
@InProceedings{saric2026phenoleafts,
author = {Sari\'c, Rijad and Azam, Basim and Khan, Sarmad and \v{C}ustovi\'c, Edhem},
title = {{PhenoLeaf-TS}: A Time-Series Benchmark for Leaf Instance Segmentation, Tracking, and Growth Stage Classification},
booktitle = {Computer Vision -- ECCV 2026},
year = {2026},
publisher = {Springer Nature Switzerland},
}