parotid gland tumours dataset
Description
Datasets consist of total of 2,110 axial MRI images, collect from 332 patients including 164 patients with normal gland ,147 patients with BATs and 21 patients with MPGTs. Axial data for both T1-inhanced and T2-weighted modalities were taken from 5 to 10 slices, resulting in a total of 2,110 axial MR images across both modalities. Organized in a hierarchical structure, the dataset features a main directory containing subdirectories, each representing a distinct category of images. Each category contains three types of files: • Images: These are the raw MRI scans that serve as the primary input for the model. The images are typically in a standard format (e.g., PNG or JPEG) and are pre-processed to ensure uniformity in size and quality. • Labels: Each image has an associated label file that provides the ground truth for the segmentation task. These labels indicate the precise boundaries of the tumour regions within the images, allowing the model to learn to differentiate between tumour and non-tumour areas. • Draw Images: These are visual representations that may include annotations or highlights of the tumour regions, providing additional context for the segmentation task. A metadata excel file includes: • patient ID • TUMOURS NAMES Tumour subtype • Age • Gander
Files
Institutions
- University of DebrecenHajdú-Bihar, Debrecen