NoduLoCC: a dataset for lung nodule localization and classification challenge from chest x-ray images

Résumé fourni par la source

Lung cancer remains one of the deadliest forms of cancer, necessitating automated tools for efficient screening. Unfortunately, these tools require high-quality labeled datasets to become efficient. Our proposal is NoduLoCC, a dataset aimed to treat detection and localization of lung nodules, facilitating experimentation and development of chest X-ray based diagnosis approaches. Existing datasets related to the detection and localization tasks are discussed. Then we present our methodology which has been applied for pre-selecting relevant data, as well as data curation with respect to our targeted tasks. Finally, we explore characteristics of the resulting dataset through statistical analysis related to the data organization, content and quality. This dataset is already used by over 20 research teams tackling the proposed detection tasks through the NoduLoCC2026 challenge, showing the usefulness of our dataset for the biomedical research community.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.