You should see your images being processed. ckpt_vidcnn-g: trained with AWGN only, with standard deviation in range.ckpt_vidcnn: uses a mixture of noises, containing AWGN, realistic and low-light noise but can generate some artefacts.1: GPU (faster but will be used only if the device has enough memory).Run the denoiser: python main_ViDeNN.py -use_gpu=1 -checkpoint_dir=ckpt_videnn -save_dir='/path/to/my/denoised_images' -test_dir='/path/to/my/images/' Using ffmpeg: ffmpeg -nostats -loglevel 0 -i /path/to/my/video /path/to/my/images/%04d.png.Video DenoisingĪs the denoiser works only with image sequences, you must export them into a directory first. If you intend to use your graphic card (GPU) to make the process faster, don't forget to install the related API (e.g CUDA for NVIDIA devices). Using a Python version under or equal to 3.6 ( Open a terminal to the downloaded directory.Clone or download and uncompress this repo.IMPORTANT! If you want to denoise data affected by Gaussian noise (AWGN), it has to be clipped between 0 and 255 before denoising it, otherwise you will get strange artifacts in your denoised output. ViDeNN works in blind conditions, it does not require any information over the content of the input noisy video. The latter has been tested only on one particular camera raw data, so it might not work on different sources. With this pretrained tensorflow model you will be able to denoise videos affected by different types of degradation, such as Additive White Gaussian Noise and videos in Low-Light conditions. This repository contains my master thesis project called ViDeNN - Deep Blind Video Denoising.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |