OpenDemoDescription
The Hair Removal Benchmark is a collection of dermoscopic images of skin lesions with and without added hair. The latter are original dermoscopic samples that contain no hair, whereas the former have been obtained by artificially generating hair, similar to that found in other dermoscopic images. In particular, we considered three hair generation mechanisms:
- Cenerative adversarial networks, by Attia et. al. (Attia, Mohamed, et al. "Realistic hair simulator for skin lesion images: A novel benchemarking tool." Artificial Intelligence in Medicine 108 (2020): 101933).
- HairSim, by Mirzaalian et. al. (Mirzaalian, Hengameh, Tim K. Lee, and Ghassan Hamarneh. "Hair enhancement in dermoscopic images using dual-channel quaternion tubularness filters and MRF-based multilabel optimization." IEEE Transactions on Image Processing 23.12 (2014): 5486-5496).
- Extracting hair masks by an automated method by Xie et al. (Xie, Feng-Ying, et al. "PDE-based unsupervised repair of hair-occluded information in dermoscopy images of melanoma." Computerized Medical Imaging and Graphics 33.4 (2009): 275-282).
The benchmark includes 618 pairs of samples with and without hair. More information about the dataset can be found in:
[1] Talavera-Martinez, Lidia, Pedro Bibiloni, and Manuel Gonzalez-Hidalgo. "Hair segmentation and removal in dermoscopic images using deep learning." IEEE Access 9 (2020): 2694-2704.
Samples
This section contains some examples of the pairs of images that are part of the database.
- Sample A
- Sample B
- Sample C
- Sample D
Access
Access to the database is free but restricted, to allow us to keep track of how much it is used, and in which projects. To get free access to the database, please contact us by email:
Acknowledgements
This work was partially supported by the R+D+i Project PID2020-113870GB-I00-“Desarrollo de herramientas de Soft Computing para la Ayuda al Diagnóstico Clínico y a la Gestión de Emergencias (HESOCODICE)”, funded by MCIN/AEI/10.13039/501100011033/.
