File size: 1,867 Bytes
3ebed6d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
---
license: cc0-1.0
tags:
  - BrainTumor
models:
  - AIOmarRehan/Brain_Tumor_Classification_with_Grad-CAM
---

# Brain Tumor MRI Dataset

## Overview
This dataset contains **7,023 human brain MRI images** organized into **four categories**:  

- Glioma  
- Meningioma  
- Pituitary  
- No Tumor  

It is designed for research in **brain tumor detection and classification** using deep learning methods. The dataset is a combination of three sources: **Figshare**, **SARTAJ**, and **Br35H**. Note that the "No Tumor" class images were sourced from the Br35H dataset.

---

## Motivation
Early and accurate detection of brain tumors is critical for treatment planning and patient survival. Automated classification using deep learning, particularly **Convolutional Neural Networks (CNNs)**, can assist in identifying tumor type, location, and malignancy grade, helping medical professionals make informed decisions faster.

---

## Dataset Details
- **Total images:** 7,023  
- **Classes:** Glioma, Meningioma, Pituitary, No Tumor  
- **Image format:** MRI scans  
- **Preprocessing:** Image sizes vary; resizing and margin removal are recommended before training.  

> Note: Some glioma images in the original SARTAJ dataset were incorrectly labeled. Those images have been replaced with correct ones from Figshare.

---

## Recommended Usage
- **Classification tasks:** Predict tumor type (glioma, meningioma, pituitary, no tumor).  
- **Segmentation tasks:** Identify tumor location within the brain MRI.  
- **Multi-task models:** Detect, classify, and locate tumors using a single CNN-based model.

---

## Citation & Feedback
If you use this dataset, please consider citing it or giving feedback. Your suggestions and comments are welcome and help improve future versions.

---

## License
[CC0: Public Domain](https://creativecommons.org/publicdomain/zero/1.0/)