diff --git a/README.md b/README.md index f440cab63afe400658d4cd55a283f2fa1c1f5f89..a0f8945ea4d176acafcf3518207e126146baabe5 100644 --- a/README.md +++ b/README.md @@ -6,19 +6,43 @@ During theinference, generation time steps are set to T = 500. The naturalizatio The implementation of CVDM can be found on the GitHub page (G. Della Maggiora, L. A. Croquevielle, N. Deshpande, H. Horsley, T. Heinis, A. Yaki-movich, Conditional variational diffusion models, https://github.com/casus/cvdm)2023. The Naturalness Factor information and documentation (including a guide on installation of the Mosaic Suite plugin) can be found in MOSAIC group, MosaicSuite documentation, https://sbalzarinilab.org/MosaicSuiteDoc/index.html. -##How to use the images? +**How to use the images?** Every folder contains 100 TIFF images representing four biological structures from the **BioSR dataset**, 25 images for each structure. -- **CCP** (Clathrin-Coated Pits) -- **ER** (Endoplasmic Reticulum) -- **F-actin** (Filamentous Actin) -- **MT** (Microtubules) +- **CCP** (Clathrin-Coated Pits) from image_1 to image_25 +- **ER** (Endoplasmic Reticulum) from image_26 to image_50 +- **F-actin** (Filamentous Actin) from image_51 to image_75 +- **MT** (Microtubules) from image_76 to image_100 The **BioSR_low_resolution_images** folder contains low-resolution images. + The **BioSR_high_resolution_images** folder contains the high-resolution grond truth images. -The **naturalized_BioSR_high_resolution_images** folder contains the naturalized high-resolution grond truth images, which are obtained by naturalize the images in **BioSR_high_resolution_images** using ImageJ Mosaic Suite plugin. -The **CVDM_inference** folder contains CVDM generated images during the inference time with images from **BioSR_low_resolution_images** as input. + +The **naturalized_BioSR_high_resolution_images** folder contains the naturalized high-resolution ground truth images, which are obtained by naturalizing the images in **BioSR_high_resolution_images** using ImageJ Mosaic Suite plugin. + +The **CVDM_inference** folder contains CVDM generated images during inference, using images from **BioSR_low_resolution_images** as input. + +The **CVDM+N_inference** folder contains images obtained by natrualizing all images in the **CVDM_inference** folder using ImageJ Mosaic Suite plugin. + +The **NCVDM_inference** folder contains images generated by NCVDM model. + +**How to use the metric code?** + +To calculate **CVDM experiment metrics**: + +- Use images from the **BioSR_high_resolution_images** folder and the **CVDM_inference** folder as pairs to compute **MS-SSIM, MAE, and PSNR**. +- Use images from the **CVDM_inference** folder to compute **RMS Contrast** and **Naturalness Factor**. + +To calculate **CVDM+N experiment metrics**: + +- Use images from the **naturalized_BioSR_high_resolution_images** folder and the **CVDM+N_inference** folder as pairs to compute **MS-SSIM, MAE, and PSNR**. +- Use images from the **CVDM+N_inference** folder to compute **RMS Contrast** and **Naturalness Factor**. + +To calculate **NCVDM experiment metrics**: + +- Use images from the **naturalized_BioSR_high_resolution_images** folder and the **NCVDM_inference** folder as pairs to compute **MS-SSIM, MAE, and PSNR**. +- Use images from the **NCVDM_inference** folder to compute **RMS Contrast** and **Naturalness Factor**.