Bootstrap

train multiple channel data using mmdetection

1. loading

(1) /anaconda3/envs/openmmlab/lib/python3.11/site-packages/mmcv/transforms/loading.py

(2) /mmdetection/mmdet/datasets/transforms/loading.py   

2. data preprocess

# channel transform

(1) /anaconda3/envs/openmmlab/lib/python3.11/site-packages/mmengine/model/base_model/data_preprocessor.py

3. Backbone network in_channel

4. Visualization:

/mmdetection/mmdet/apis/det_inferencer.py

/mmdetection/mmdet/visualization/local_visualizer.py

;