Summary of medical image open source datasets (2)

Edit 丨 Extreme Market Platform

Blood Cell Image Dataset

Dataset download address:

This dataset contains 12,500 enhanced blood cell images (JPEG) with cell type labels (CSV). Each of the 4 different cell types has about 3,000 images grouped into 4 different folders (according to the cell type). The cell types are eosinophils, lymphocytes, monocytes, and neutrophils.

Brain Tumor MRI Dataset

Dataset download address:

This dataset contains 7022 human brain MRI images, divided into 4 categories: glioma-meningioma-no tumor and pituitary. Note that the sizes of the images in this dataset are different. You can resize the image to your desired size after preprocessing and removing extra margins.

OCT image dataset of diabetic macular edema

Dataset download address:

Patients in Duke Eye Center's Retina Medicine Practice were retroactively identified using the Duke Enterprise Data Unified Content Explorer search engine and used the DME (ICD-9 362.07) billing codes associated with their visits. An ophthalmologist then identified 6 clinically imaged patients with severe DME pathology and variable image quality using the standard Spectralis (Heidelberg Engineering, Heidelberg, Germany) 61-line volumetric scanning protocol.

X-ray Image Dataset of Body Parts

Dataset download address:

This dataset collects X-ray images from various parts of the body

Eye Disease Deep Learning Dataset

Dataset download address:

712 eye stain images of flaky corneal ulcers

Dermatology Dataset

Dataset download address:

The image data of 23 types of skin diseases, the total number of images is about 19,500, of which about 15,500 have been segmented in the training set, and the rest are segmented in the test set.

Heart Attack Analysis and Prediction Dataset

Dataset download address:

This dataset contains data of 303 cardiac patients.

Knee X-ray Image Dataset

Dataset download address:

The dataset consists of 1650 electron X-ray images of knee joints collected from well-known hospitals and diagnostic centers. X-ray images were acquired using a PROTEC PRS 500E X-ray machine. The original image is an 8-bit grayscale image. Each X-ray knee X-ray image was manually annotated/labeled by 2 medical experts according to Kellgren and Lawrence grades.

Guess you like