- ✅ Release the final test result
- ✅ Release the processing code
- ✅ Release the second finetune process
- ✅ Update the final model weights 0108
- [ ] Release the source code
- [ ] Release the first-stage training process
Create a conda environment & Install requirments
conda create -n motiondissolve python==3.10.16
conda activate motiondissolve
pip install -r requirements.txt
You can leverage processing_event.py on the original event data to obtain a fixed frame event input.
python output3_test_guass.py- running output3_test_guass.py to obtain the final output ./results/bin. You can also download our final results straightly from ./results/bin.
python output3_test_image_time.py- running output3_test_image_time.py to obtain an inference time of our methods.
python bin2rgb.py- running bin2rgb.py to obtain visualization of our output
Thanks to the inspiration and codes from AHDINet, thanks for their generous open source.