TPDNet

Count-Supervised Network (CSNet) can capture the 3D information of objects from a single RGB image for fruits and vegetables in fields.

paper: TPDNet: Triple phenotype deepen networks for Monocular 3D Object Detection of Melons and Fruits in Fields

The Overview of CSNet

About Data

We use the Wax Gourd 3D Object Detection DataSets (dataset).

Training

./launcher/eval.sh config/config.py

Testing

./launcher/eval.sh config/config.py

Result