Faheem Ashraf

Academic

Career

Verbatim text

Image Inpainting Using Deep Learning Models 
A multi-platform-based app is being which is used to denoise, colourise B&W images along with 
super resolution and restoration of images. The app is using deep learning models at back-end to 
perform the following actions. There are two interfaces, web interface where user can select what 
he wants to do with the image and then upload the image and let it process. After process is 
complete, the user will be displayed with the processed image which he can save by right clicking 
on it and hit save image. The second interface is a mobile app where user can select the action he 
wants to perform and then given option to give image from camera or select from gallery, after 
which the image is uploaded on server and processed and user can view the processed image on 
the final screen. The backend is deployed on cloud to ensure stability and since it requires high 
computational 
power, 
cloud 
is 
the 
best 
option. 
Features include: 
 Multi-platform app 
 Denoise Images 
 Colourise black and white images 
 Super resolution of pixelated images 
 Restore damaged images. 
Take live pictures from mobile camera and process 
 
 
 
 
 
 
 
 
 
 
 
 
Technology Used: 
Python, Flask, Flutter, GitHub, TensorFlow, Keras 
Supervisor Name: 
Dr. Irum Inayat 
Group Members:   
Faheem Ashraf (17I-0061) 
Muhammad Mehlab (17I-0188) 
Talha Muneer (17I-0229)