Sadbhawna

I'm working as an Assistant Professor in the CSE Department at MNIT Jaipur since January 2024. Before that I was an Institute Post Doctoral Fellow at IIT Madras. I recieved my Ph.D. from the CSE Department at IIT Jammu advised by Dr. Vinit Jakhetiya in January 2023. I primarily work in the area of Image Processing, Computer Vison and Machine Learning. My main focus includes analyzing and enhancing the perceptual quality of super-resolved images/videos, 3D synthesized views.

I completed my M.Tech. from SLIET Longowal in 2018 and B.Tech from HPTU in 2016.

Email  /  CV  /  Google Scholar   /  Github  /  LinkedIn  /  Twitter  / 

19/02/2024- One paper has been accepted in IEEE Transactions on Multimedia.

11/01/2024- I have joined MNIT Jaipur as an Assistant Professor.

16/04/2023- I have joined IIT Madras as an Institute Post Doctoral Fellow.

05/04/2023- One paper has been accepted in CVPR Workshops 2023 (NTIRE: New trends in image restoration and enhancement).

07/01/2023- I successfully defended my Ph.D. Thesis.

04/09/2022- One paper has been accepted in IEEE Transactions on Multimedia.

15/03/2022- Started working with Dr. Sunil Jaiswal, Head R&D, K|Lens GmbH

25/12/2021- Selected as PIEF candidate by IGSTC for a 6 months internship in Germany. (Results Link)

30/11/2021- One paper has been accepted in AAAI Student Abstract Program.

04/10/2021- Two papers has been accepted in IEEE Transactions on Image Processing.

15/06/2021- Our team has been awarded 600 USD sponsored by Google in ICASSP 2021 SPGC Grand Challenge on being Runner Ups.

Spring 2024- CST310 Computer Graphics Course Website

Spring 2024- CS Computer Organization and Architecture Course Website

This work shows how varied random degradations can contribute to learning an effective VSR model, especially for real-world video artifacts.

In this work, we propose a new and efficient quality assessment algorithm based upon the variation in the depth of 3D synthesized and reference views.

We created a test dataset to analyze the need for a new perceptual metric for 3D synthesized views.

In this work, we extract the perceptually important deep features from the pre-trained VGG-16 architectures on the Laplacian pyramid to predict the quality of 3D synthesized views.

We proposed a Convolutional Neural Network (CNN) based algorithm that identifies the blocks with stretching artifacts and further incorporates the number of blocks with the stretching artifacts to predict the quality of 3D-synthesized views.

We proposed a two-stage Convolutional Neural Network (CNN) based classification framework for detecting COVID-19 and Community Acquired Pneumonia (CAP) using the chest Computed Tomography (CT) scan images

We proposed a perceptual metric for 3D views based on the difference in propoerties of synthetic and natural images.