IBSU-1432 dataset on videoendoscopy frame quality classification for laryngoscopy (NBI modality)

How to cite

Nogal, P., Buchwald, M., Staśkiewicz, M., Kupiński, S., Pukacki, J., Mazurek, C., ... & Wierzbicka, M. (2022). Endoluminal larynx anatomy model–towards facilitating deep learning and defining standards for medical images evaluation with artificial intelligence algorithms. Polish Journal of Otolaryngology, 76(5), 37-45. https://doi.org/10.5604/01.3001.0015.9501

Paderno, A., Piazza, C., Del Bon, F., Lancini, D., Tanagli, S., Deganello, A., … Moccia, S. (2021). Deep Learning for Automatic Segmentation of Oral and Oropharyngeal Cancer Using Narrow Band Imaging: Preliminary Experience in a Clinical Perspective. Frontiers in Oncology, 11(March), 1–12. https://doi.org/10.3389/fonc.2021.626602

Moccia, S., Vanone, G. O., Momi, E. De, Laborai, A., Guastini, L., Peretti, G., & Mattos, L. S. (2018). Learning-based classification of informative laryngoscopic frames. Computer Methods and Programs in Biomedicine, 158, 21–30. https://doi.org/10.1016/j.cmpb.2018.01.030

Description

The presented dataset consists of 1432 laryngeal endoscopy frames of different acquisition quality. The four classes were distinguished (after Moccia et al., 2018):

1. Informative frames (436),
2. Blurred (383),
3. Saliva/specular reflections (321), and
4. Underexposed frames (292).

(In total, 1432 = 436 I + 383 B + 321 S + 292 U.)

Acknowledgements

Alberto Paderno, MD PhD – Brescia data part

Sara Moccia, PhD – IBSU-720 frames dataset from Zenodo: https://zenodo.org/record/1162784#.Ycrmfi1Q1qt

Małgorzata Wierzbicka, MD PhD – Poznan data part

Piotr Nogal, MD – Poznan data part

Joanna Jackowska, MD PhD – Poznan data part

Hanna Klimza, MD PhD – Poznan data part