diff --git a/config/bih/landingPage.json b/config/bih/landingPage.json index aa98fa3..48df493 100644 --- a/config/bih/landingPage.json +++ b/config/bih/landingPage.json @@ -23,7 +23,13 @@ "text": "Queries and subsequent analyses in the BIH run across structured data aggregated from all the independent data nodes connected through the data fabric." }, { - "text": "Data files are not stored in the BIH. Instead, BIH records contain unique identifiers for files that researchers can use to access files from the connected data nodes in which they’re stored." + "text": "Data files are not stored in the BIH. Instead, BIH records contain unique identifiers for files that researchers can use to access files from the connected data nodes in which they're stored." + }, + { + "text": "Important note for users: Due to the nature of federated data systems like this one, it is possible for certain datasets or images to appear as duplicates across multiple data nodes. To ensure the integrity and reliability of your research, we strongly recommend exercising due diligence in identifying and handling duplicate images in any datasets you select. This is particularly crucial when preparing data for use in training AI models, as duplicates in training data can introduce biases, overfitting, and other issues that may compromise the performance and generalizability of your models." + }, + { + "text": "We recommend users check for duplicated images, using unique IDs like Imaging Study and Series UIDs, perform thorough data validation before proceeding with any analysis or AI model training, and document your process by keeping detailed records of how duplicates were identified and handled to enhance the reproducibility of your research." } ], "right": [