"An Annotated Clinical Image Dataset for Deep Learning-Based Classification of Oral Lesions"

Overview:

This intraoral dataset is collected using mobile phones and DSLR camera. It comprises 9,201 intraoral images, which are subdivided according to the presence and type of oral lesion. These include 4,405 images classified as ‘Normal,’ 2,314 as ‘Low Risk of malignant transformation,’ and 2,482 as ‘High Risk of malignant transformation'. The dataset consisted solely of original images, without any augmented photos. The dataset was comprehensively annotated by Oral Medicine specialists using LabelMe.exe in JSON (JavaScript Object Notation) format. The dataset contains the following files:

• normal.zip : intraoral annotated images in the 'normal' category.
• Low.zip: intraoral annotated images in the 'Low Risk of malignant transformation' category.
• High.zip: intraoral annotated images in the 'High Risk of malignant transformation' category.

Access Request Guidelines

To request access to this dataset, you must comply with the following terms and conditions:

1. Submission Requirements:

   • Requests must be made through a Zenodo account linked to an official email address associated with an active and recognized institution, organization, or research entity (e.g., *.edu, *.gov, *.org).
   • Generic email domains (e.g., Gmail, Yahoo) will not be accepted under any circumstances.

2. Required Information:
   The request message must include:

   • A concise description (maximum 250 words) outlining the intended use of the dataset, specifying the purpose, methodology, and scope of the study.
   • The full name and institutional affiliation of the Principal Investigator supervising the research.
   • A direct link to the Principal Investigator’s profile on the official website of their affiliated institution.

3. Non-Commercial Use Only:

   • The dataset is strictly limited to non-commercial use, such as academic research, educational activities, or other purposes without monetary or commercial intent.
   • Commercial use, including direct or indirect financial gain (e.g., incorporation into paid services, monetized products, or for-profit applications), is strictly prohibited.

4. Restrictions on Modifications and Derivatives:

   • The dataset must be used in its original form. Modifications, derivatives, or adaptations, such as altering the structure, creating augmented data, or redistributing transformed versions, are not permitted.
   • Preprocessing for research purposes (e.g., resizing, normalization) is allowed but must not result in a redistributable version of the dataset.

5. Citation Requirement:

   • Proper attribution to the dataset creator is mandatory. All publications or presentations based on this dataset must cite the original article, following the citation format provided with the dataset.

6. Review and Processing:

   • Complete and compliant requests will be reviewed within 3-5 business days. Incomplete or non-compliant requests will require resubmission.

7. Ethical Compliance:

   • Users must adhere to all relevant ethical guidelines, institutional approvals, and privacy regulations applicable to the dataset's use.