{"id":20373507,"url":"https://github.com/megengine/pmrid","last_synced_at":"2025-04-13T08:27:15.113Z","repository":{"id":41390724,"uuid":"313955135","full_name":"MegEngine/PMRID","owner":"MegEngine","description":"ECCV2020 - Practical Deep Raw Image Denoising on Mobile Devices","archived":false,"fork":false,"pushed_at":"2023-04-18T10:22:47.000Z","size":7532,"stargazers_count":237,"open_issues_count":17,"forks_count":39,"subscribers_count":8,"default_branch":"main","last_synced_at":"2025-03-27T00:12:00.215Z","etag":null,"topics":["denoise","megengine"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/MegEngine.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2020-11-18T14:12:01.000Z","updated_at":"2025-03-26T14:14:39.000Z","dependencies_parsed_at":"2022-09-03T17:50:38.681Z","dependency_job_id":"d8271c01-0d7a-4c2e-a6d7-ff9423352484","html_url":"https://github.com/MegEngine/PMRID","commit_stats":null,"previous_names":["megvii-research/pmrid"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MegEngine%2FPMRID","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MegEngine%2FPMRID/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MegEngine%2FPMRID/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MegEngine%2FPMRID/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/MegEngine","download_url":"https://codeload.github.com/MegEngine/PMRID/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248683265,"owners_count":21144877,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["denoise","megengine"],"created_at":"2024-11-15T01:18:42.046Z","updated_at":"2025-04-13T08:27:15.087Z","avatar_url":"https://github.com/MegEngine.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Practical Mobile Raw Image Denoising (PMRID)\n\nCode and dataset for ECCV20 paper [Practical Deep Raw Image Denoising on Mobile Devices](https://arxiv.org/abs/2010.06935).\n\n## Dataset\n\n### Downloads\n- [OneDrive](https://megvii-my.sharepoint.cn/:f:/g/personal/wangyuzhi_megvii_com/Et4v2Z7CkRxHnbcFUq6RXZMBfXUrlm_Se5OVDvcdujVsMA?e=vcfJWs)\n- [Kaggle](https://www.kaggle.com/dataset/1bdc5cd707cfbb3ee842eb3cbfe93495dbba88017d29f295f8edbcb8f8790556)\n\n### Usage\n\nThe dataset includes two 7zip files:\n- `reno10x_noise.7z` contains DNG raw images shot by an _OPPO Reno 10x_ phone for noise parameter estimation (refer Sec 3.1 and 5.1 in the paper)\n- `PMRID.7z` is the benchmark dataset described in Sec 5.2 in the paper\n\nThe structure of `PMRID.7z` is\n```\n- benchmark.json  # meta info\n- Scene1/\n  \\- Bright/\n     \\- exposure-case1/ \n         \\- input.raw   # RAW data for noisy image in uint16\n          - gt.raw      # RAW data for clean image in uint16\n      + case2/\n  + Dark/\n+ Secne2/\n```\n\nAll metadata for images are listed in `benchmark.json`:\n```python\n{\n   \"input\": \"path/to/noisy_input.raw\",\n   \"gt\": \"path/to/clean_gt.raw\",\n   \"meta\": {\n       \"name\": \"case_name\",\n       \"scene_id\": \"scene_name\",\n       \"light\": \"light condition\",\n       \"ISO\": \"ISO\",\n       \"exp_time\": \"exposure time\",\n       \"bayer_pattern\": \"BGGR\",\n       \"shape\": [3000, 4000],\n       \"wb_gain\": [r_gain, g_gain, b_gain],\n       \"CCM\": [   # 3x3 color correction matrix\n           [c11, c12, c13], \n           [c21, c22, c23], \n           [c31, c32, c33]\n       ],\n       \"ROIs\": [  # patch ROIs to calculate PSNR and SSIM, x0 is topleft\n           [topleft_w, topleft_h, bottomright_w, bottomright_h]\n       ]\n   }\n}\n```\n\n## Pre-trained Models and Benchmark Script\n\nBoth [PyTorch](https://pytorch.org/) and [MegEngine](https://megengine.org.cn/) pre-trained models are provided in the `models` directory. \nThe benchmark script is written for models trained with MegEngine. `Python \u003e= 3.6` is required to run the benchmark script.\n\n```\npip install -r requirements.txt\npython3 run_benchmark.py --benchmark /path/to/PMRID/benchmark.json models/mge_pretrained.ckp\n```\n\n\n## Citation\n```\n@inproceedings{wang2020,\n\ttitle={Practical Deep Raw Image Denoising on Mobile Devices},\n\tauthor={Wang, Yuzhi and Huang, Haibin and Xu, Qin and Liu, Jiaming and Liu, Yiqun and Wang, Jue},\n\tbooktitle={European Conference on Computer Vision (ECCV)},\n\tyear={2020},\n\tpages={1--16}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmegengine%2Fpmrid","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmegengine%2Fpmrid","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmegengine%2Fpmrid/lists"}