Densenet Tumor Model
This repository hosts the contributor source files for the Densenet Tumor model.
Metadata
| Application Area | Problem Type | Data Type | Source |
|---|---|---|---|
| Brain MRI(Magnetic Resonance Imaging) | Classification | JPG | Brain Tumor |
Model
| Description | Provenance | Architecture | Learning_type |
|---|---|---|---|
| This is a PyTorch implementation of the DenseNet-121 architecture. | Contributed by author | A three-dimensional (3D) convolutional neural network (CNN) | Transfer Learning |
About Dataset
Abstract
Brain tumors are among the most aggressive diseases affecting both children and adults. They account for 85 to 90 percent of all primary Central Nervous System (CNS) tumors. Each year, around 11,700 people are diagnosed with a brain tumor. The 5-year survival rate for individuals with a cancerous brain or CNS tumor is approximately 34 percent for men and 36 percent for women. Brain tumors can be classified as Benign Tumors, Malignant Tumors, Pituitary Tumors, and others. Accurate treatment planning and diagnostics are crucial to improving patient life expectancy. The most effective method for detecting brain tumors is Magnetic Resonance Imaging (MRI). However, the vast amount of image data generated from MRI scans requires detailed examination by radiologists, and manual analysis can be error-prone due to the complexities involved in brain tumors and their characteristics.
Automated classification techniques using Machine Learning (ML) and Artificial Intelligence (AI) have consistently demonstrated higher accuracy than manual classification. Therefore, implementing a system that utilizes Deep Learning Algorithms such as Convolutional Neural Networks (CNN), Artificial Neural Networks (ANN), and Transfer Learning (TL) could provide valuable assistance to medical professionals worldwide.
Context
Brain tumors are complex, with significant variations in size, location, and characteristics. These variations make it challenging to fully understand the tumor's nature. MRI analysis requires the expertise of professional neurosurgeons. In many developing countries, the lack of skilled doctors and limited knowledge about brain tumors make it particularly challenging and time-consuming to generate accurate MRI reports. An automated, cloud-based system could address these issues by providing accurate tumor classification and analysis, making MRI evaluations more efficient and accessible.