Medical Image Analysis Brain Tumor Segmentation

This research is focused on detecting gliomas, the most agressive brain tumor using Deep Convolutional Neural Networks (CNN). Datasets are from BRATS challenge (https://www.smir.ch/BRATS/).



Ischemic Stroke Lesion Segmentation
This research is focused on segmenting lesions caused by ischemic strokes. Like the previous research, CNNs are used here. Datasets are from ISLES challenges (https://www.smir.ch/ISLES/).
       
Automatic License Plate Recognition (ALPR)
ALPR for Iranian Vehicles
In this research, plate area and details are automatically extracted from Iranian license plates to help with automatic opening of entrance gates and issuing tickets for traffic violations.
       

Security Detecting FDI Attacks in Power Grids
False Data Injection (FDI) attacks bypass conventional Bad Data Detection (BDD) methods. This research is focused on using Deep Learning methods such as Long Short Term Memroy (LSTM) networks for detecting FDI.


Detecting Android Malwares
Dynamic properties of Android executables have been less considered in Android malware detection researches. We aim to use Deep Learning methods so for this.
       

Linguistics Handwritten Words Recognition
Persian handwritten words are difficult for automatic recognition. In this research, widely used CNN architectures such as AlexNet and GoogLeNet are employed for this.