Welcome to Evently

Lorem ipsum proin gravida nibh vel veali quetean sollic lorem quis bibendum nibh vel velit.


Simply enter your keyword and we will help you find what you need.

What are you looking for?

casibom güncel giriş rokubet giriş tümbet kavbet lunabet hitbet dinamobet baywin betine betorspin bayspin baywin betmatik istanbul escort şişli escort antalya escort crazy frog horse fuck porn video japanese milf porn lesbian dildo xxx sleeping step sister

Paper ID 7811

Paper Information

Detection of Brain Tumor in 3D MRI

Paper ID: 7811

Ahtesham Shahid

Sajid Khan

Usman Habib

Sharjeel Mukhtar

Sophia Shahid

Ammar Ahmed


This is an application for the neurosurgeon to detect the different kinds of tumors from high grade to low grade in 3D MRI images taken from the BRAT 2020 dataset. In this work, the proposed system’s accuracy and performance have been improved by using weak supervision in the training system and modified SOM (self-organized mapping) [2] as compared to the watershed algorithm. The proposed model consists of four layers input layer, hidden layer 1, hidden layer 2, and an output layer performing their respective functionalities on data. The first layer is an input layer in which pre-processing on an MR image is carried out for its modification in a way it is ready to propagate to other layers. Hidden layer 1 is based on computer vision in which the Watershed algorithm is applied. Hidden layer 2 consists of SOM which is a machine learning algorithm providing clustering of data of MR image and passing data to SVM as a weak label along with the process of feature extraction for classifier SVM. The final layer is an output layer identifying tumor regions along with categorizing tumors in that MR image to find the tumor regions and results have achieved higher accuracy along with the performance of the system by reducing the training time. Three models were trained, tested, and evaluated named SVM [4], SVM with Gaussian Kernel, and KNN. Our model SVM with Gaussian Kernel gives the highest accuracy of 89% concerning simple SVM and KNN which gave 85% and 78% accuracy respectively on BRAts 2020 Dataset.