PH² database
| 1 - Introduction |
The increasing incidence of melanoma has recently promoted the development of computer-aided diagnosis systems for the classification of dermoscopic images. The PH² dataset has been developed for research and benchmarking purposes, in order to facilitate comparative studies on both segmentation and classification algorithms of dermoscopic images. PH² is a dermoscopic image database acquired at the Dermatology Service of Hospital Pedro Hispano, Matosinhos, Portugal.
| 2 - Data Description |
The dermoscopic images were obtained at the Dermatology Service of Hospital Pedro Hispano (Matosinhos, Portugal) under the same conditions through Tuebinger Mole Analyzer system using a magnification of 20x. They are 8-bit RGB color images with a resolution of 768x560 pixels.
This image database contains a total of 200 dermoscopic images of melanocytic lesions, including 80 common nevi, 80 atypical nevi, and 40 melanomas. The PH² database includes medical annotation of all the images namely medical segmentation of the lesion, clinical and histological diagnosis and the assessment of several dermoscopic criteria (colors; pigment network; dots/globules; streaks; regression areas; blue-whitish veil).
The assessment of each parameter was performed by an expert dermatologist, according to the following parameters:
| Criterion | PH² Segmentation |
| Clinical Diagnosis | 0 - Common Nevus |
| 1 - Atypical Nevus | |
| 2 - Melanoma | |
| Lesion Segmentation | Available as a binary mask (with the same size of the original image). |
| Color Segmentation | Available as a binary mask (with the same size of the original image) (If available). |
| Asymmetry | 0 - Fully Simmetry |
| 1 – Asymmetry in One Axis | |
| 2 - Fully Asimmetry | |
| Pigment Network | AT - Atypical |
| T - Typical | |
| Dots/Globules | A - Absent |
| AT - Atypical | |
| T - Typical | |
| Streaks | A - Absent |
| P - Present | |
| Regression Areas | A - Absent |
| P - Present | |
| Blue Whitish Veil | A - Absent |
| P - Present | |
| Colors | 1 - White |
| 2 - Red | |
| 3 - Light-Brown | |
| 4 - Dark-Brown | |
| 5 - Blue-Gray | |
| 6 - Black |
| 3 - Terms of Use |
The data included in the PH² database can be used for research and educational purposes. It is important to note that redistribution and commercial use is not allowed. All publications that make use of this dataset must cite the following paper:
| 4 - PH² Browser: a PH² Database Related Tool |
4.1 - About
The PH² database can be used independently by the user or with the PH² Browser tool.
The PH² Browser software is a customized tool of the PH² image database. This tool allows the visualization of the medical segmentations and classifications of each image of the database. Moreover, it allows searching dermoscopic images according to the desired criteria as well as the exportation of the selected images, along with the corresponding segmentation and classification, into several formats (See tutorial for more information).
4.2 - Installation
The PH² Browser tool was developed in Matlab™ and is available for public as an MS Windows™ (x86/x64) executable.
To install the PH² Browser software, it is necessary to download the entire PH² winrar package. After downloading this package the user must unzip the “PH² Browser” folder. Inside this folder the user can find a MS Windows™ (x86) version as well as a MS Windows™ (x64) version of the software. Then, the user should select one of these versions accordingly to the Windows version of the computer in which the software will be used. If MATLAB™ is already installed in the user’s computer, no other installation steps are required. Otherwise, in a computer in which MATLAB™ is not installed, the PH² Browser tool can also be used after a simple installation of the MATLAB™ Compiler Runtime (MCR).
- Links to download: MRC Installer (x86) / MRC Installer (x64)
After all installation steps are completed the user must double-click on the PH²_Browser.exe file to start using the software.
| 5 - Download |
The PH² Database pack (image database and medical annotation) and the PH² Browser can be downloaded here after you fill a quick registration form.
- PH² Dataset - download link
- PH² Browser - download link
| 6 - Benchmarking results |
The following table presents the classification algorithms, along with their results, that have been applied to the PH² database for benchmarking purposes.
| Authors/Paper | Extracted features | Sensibility/Specificity | |
| C.Barata, M.Ruela, et al., Two Systems for the Detection of Melanomas in Dermoscopy Images using Texture and Color Features, IEEE Systems Journal, no. 99, pp. 1-15, 2013. | Global Method | Color features | 90% / 89% |
| Texture features | 93% / 78% | ||
| Local method | Color features | 93% / 84% | |
| Texture features | 88% / 76% | ||
ADD YOUR RESULTS
We encourage all researchers to run their algorithms in the PH² database and, then, send us the results. To ensure the integrity of results, you should apply your classification algorithms in all images of the PH² database. In addition, you should indicate a paper, a report, or even a website in which your algorithm is described.
To upload your results, contact Dr. Terera Mendonça by e-mail: tmendo@fc.up.pt.
| 7 - Support |
The PH² Database was built up through a joint research collaboration between the Universidade do Porto, Instituto Superior Técnico Lisboa, and the Dermatology Service of Hospital Pedro Hispano in Matosinhos, Portugal. This research work was supported by the Portuguese Foundation for Science and Technology (FCT) through the projects PTDC/SAU-BEB/103471/2008 and PEst-OE/EEI/LA0009/2011, by FEDER funds through the Operational Programme Factors of Competitiveness (COMPETE), by Portuguese funds through the Center for Research and Development in Mathematics and Applications and the FCT, within project PEst-C/MAT/UI4106/2011 with COMPETE number FCOMP-01-0124-FEDER-022690.
Next Events
Computer vision and Machine learning events:
Dermoscopy events:
Visitors
ADDI Work
- Ongoing Work- Completed Work