I am currently working on reproducing the downstream classification experiment in the RetCCL paper. I want to ask you two questions:
Q1: Could you provide me more details about your setup, such as the number of FC layers, activation function, hyperparameters, preprocessing steps before feeding them to the model, etc.?
Q2: According to your provided URL, I need help finding the DiagSet-A.2 dataset (https://ai-econsilio.diag.pl/). The datasets that I found are DiagSet-A, DiagSet-B, and DiagSet-C. However, in detail, no dataset consists of 4 classes (Normal, Gleason score 3, Gleason score 4, and Gleason score 5) paired with an image patch (224 x 224 px). Please advise me on which one your DiagSet-A.2 dataset is.
These will help me reproduce the results a lot.
Thank you in advance for your contribution.
Dear, RetCCL's authors
I am currently working on reproducing the downstream classification experiment in the RetCCL paper. I want to ask you two questions: Q1: Could you provide me more details about your setup, such as the number of FC layers, activation function, hyperparameters, preprocessing steps before feeding them to the model, etc.? Q2: According to your provided URL, I need help finding the DiagSet-A.2 dataset (https://ai-econsilio.diag.pl/). The datasets that I found are DiagSet-A, DiagSet-B, and DiagSet-C. However, in detail, no dataset consists of 4 classes (Normal, Gleason score 3, Gleason score 4, and Gleason score 5) paired with an image patch (224 x 224 px). Please advise me on which one your DiagSet-A.2 dataset is.
These will help me reproduce the results a lot. Thank you in advance for your contribution.