RichardObi / medigan

medigan - A Python Library of Pretrained Generative Models for Medical Image Synthesis
https://medigan.readthedocs.io/en/latest/
MIT License
133 stars 14 forks source link

Model Integration Request for medigan: 00022_WGAN_CARDIAC_AGING #54

Closed faildeny closed 1 year ago

faildeny commented 1 year ago

Creator: Víctor M. Campello Affiliation: University of Barcelona Stored in: https://zenodo.org/record/7446930

Model Metadata:


   "00022_WGAN_CARDIAC_AGING": {
      "execution": {
         "package_name": "00022_WGAN_CARDIAC_AGING",
         "package_link": "https://zenodo.org/record/7446930",
         "model_name": "model",
         "extension": ".ckpt",
         "image_size": [
            256,
            256
         ],
         "dependencies": [
            "nibabel==3.2.1",
            "pytorch-lightning==1.4.7",
            "pandas",
            "comet-ml",
            "monai",
            "grad-cam",
            "matplotlib",
            "monai[skimage]",
            "munch==2.5.0",
            "pillow==7.0.0",
            "ffmpeg-python==0.2.0"
         ],
         "generate_method": {
            "name": "generate_GAN_images",
            "args": {
               "base": [
                  "model_file",
                  "output_path",
                  "save_images",
                  "num_samples"
               ],
               "custom": {
                  "image_paths_input": [
                     "models/00022_WGAN_CARDIAC_AGING/sample_image.png"
                  ],
                  "aging_input": [
                     -4
                  ],
                  "data_type": "2d",
                  "view": "la",
                  "subcat": "2ch"
               }
            }
         }
      },
      "selection": {
         "performance": {},
         "use_cases": [
            "classification",
            "segmentation"
         ],
         "organ": [
            "heart",
            "chest"
         ],
         "modality": [
            "MRI",
            "Cardiac imaging",
            "Cardiography",
            "full-field digital"
         ],
         "vendors": [],
         "centres": [],
         "function": [
            "image to image",
            "image generation",
            "data augmentation"
         ],
         "condition": [
            "age"
         ],
         "dataset": [
            "UK Biobank"
         ],
         "augmentations": [
            "resize"
         ],
         "generates": [
            "cardiac image",
            "full-field digital"
         ],
         "height": 256,
         "width": 256,
         "depth": null,
         "type": "pix2pix",
         "license": null,
         "dataset_type": "non-public",
         "privacy_preservation": null,
         "tags": [
            "Cardiac imaging",
            "pix2pix",
            "Pix2Pix"
         ],
         "year": "2022"
      },
      "description": {
         "title": "Generates cardiac images with age offset from real images (Trained on UK Biobank)",
         "provided_date": "",
         "trained_date": "",
         "provided_after_epoch": 299,
         "version": "0.0.1",
         "publication": "https://www.frontiersin.org/articles/10.3389/fcvm.2022.983091",
         "doi": [
            ""
         ],
         "inputs": [
            "input_image_paths: default=[\"models/00022_WGAN_CARDIAC_AGING/sample_image.png\"] help=List of image paths to apply aging.",
            "aging_input: default=[-4] help=List of age offset values for each image.",
            "data_type: default=\"2d\" help=",
            "view: default=['oval', 'lobulated'] help=",
            "subcat: default=5, help="
         ],
         "comment": ""
      }
   }
}
RichardObi commented 1 year ago

Addressed in PR55