Tracking script and documentation need updates

[TNodeCode]

Currently the tracking process described in the README file doesn't work. I had to do the following stept to fix the error:

Replace all occurences of args.images in the file src/spine_detection/tracking_mmdet.py by args.input. Otherwise the --input argument will not be used in the Python code.

Modify the command to track the objects in the following way:

python src/spine_detection/tracking_mmdet.py \
    --input "<path-to-files>/SR052N1D2day1stack1*.png" \
    --model Cascade_RCNN_model \
    --model_type Cascade-RCNN \
    --param_config lr_0.001_warmup_None_momentum_0.6_L2_3e-06_run_1 \
    --model_epoch epoch_10 \
    --save_images

Rename the directory Cascade_RCNN in the directory tutorial_exps to Cascade_RCNN_no_data_augmentation. You can do this by running the following command: mv tutorial_exps/Cascade_RCNN_model/ tutorial_exps/Cascade_RCNN_no_data_augmentation.

[SaILaIDiN]

Thanks for pointing out the args.images parameter! This was an artifact of a previous version. We have updated the tracking_mmdet.py file for this aspect.

The naming of the directory is just explanatory. But we will consider unifying the folder structure and naming for training, tracking, prediction and evaluation.

Best regards