The Global Wheat Detection competition was hosted by the Univ. of Saskatchewan on Kaggle. The final solutions were mostly EfficientDet and Yolo-V5 models. But we experimented with a two-stage OD model, Mask-RCNN. Recently, we have created a gradio app to infer this model instantly. You can try:
(Recommend gradio environment for quick inference on test set.)
|
|
