Discrete Differential Geometry
Mesh Denoising
Shape Analysis
Medical Imaging
Machine Learning
Computer Vision

Publications and Preprints

Mesh Denoising based on Normal Voting Tensor and Binary Optimization
A tensor multiplication based smoothing algorithm that follows a two step denoising method. Unlike other traditional averaging approaches, our approach uses an element based normal voting tensor to compute smooth surfaces.
Yadav, S.K., Reitebuch, U., Polthier, K.

Robust and Stable Mesh De-Noising. (In preparation)
In this paper, we introduce a de-noising algorithm that concerns mainly on the robustness against the high intensity of the noise and better mesh quality.
Yadav, S.K., Reitebuch, U., Polthier, K.

3D Super-Resolution using Multiple Low Resolution Scans: A Regularization Approach. (In preparation)
In this report, we are proposing a method that results in a true 3D super-resolution model rather than simple 2.5D model by merging low resolution scans of an object.
Yadav, S.K., Polthier, K.


Retinal Shape Analysis:
This is a collaboration project between Charite Medicine University and Free University of Berlin. This project mainly focuses on the shape analysis of the human retinal structure and mathematical modelling of the human retina. The main goal of this project to diagnose the different eye related disease like glaucoma, MS, IIH etcmorphometrically.

Mesh Denoising(March 2014 - August 2015)
This project was funded by Dahlem Research School (DRS) Berlin. Denosing is one of the most active and fascinating research areas in geometry processing as the digital scanning devices becoming the widespread to capture the 3d points of the surface of 3D objects. In this projects, we developed de-noising algorithms that concern mainly on the robustness against the high intensity of the noise and better mesh quality and enhance the sharp features.

Super Resolution of the 3D Scans(June 2013 – February 2014)
This project was funded by Einstein Foundation Program (EFP) Berlin. The resolution of an object surface can be increased by merging various low resolution scans measured by the laser scanner. In this project, we are proposing a method that results in a true 3D super-resolution model rather than simple 2.5D model by merging multiple low resolution scans of an object.


Master Thesis:

High Accuracy Sub-Pixel Based Correspondence Point Matching Between Two Images
Computer Vision and Remote Sensing, Technical University Berlin, Germany 2013.
In this work, we are investigating the minimum shifting between two image in terms of sub pixels, whereas images are shifted in micrometer range. We are using Area Based Matching, so computational complexity is more. Another focus of this work is to reduce the computational complexity for calculating the corresponding point between two images. To reduce computational complexity we are using 1D BLPOC (Band Limited Phase Only Correlation), in which we search for corresponding point along the epipolar line only. So this algorithm has lower computational complexity in comparison to other existing algorithm.

Conference And Workshop Attendance

International Geometry Summit (IGS) 2016 - 20 - 24 June 2016, Berlin

IGS Summer School 2016 - 18-19 June 2016, Berlin

Computer graphics international 2013 - 11-14 June 2016, Hannover

About Me

I'm a Ph.D. student in the mathematical geometry processing group headed by Prof. Dr. Konrad Polthier at Free University Berlin.

My research interests are somewhere in between Discrete Differential Geometry, Medical Imaging and Computer Vision. Currently, I am working on triangulated surface smoothing. I am also working on the morphometry of the retina of a human eye in Charite Medicine University Berlin.

I received master’s degree (M-Tech.) in Signal and Image Processing from IIT Roorkee, India. My master’s thesis was done in supervision of Prof. Dr. Olaf Hellwich Head of Computer Vision And Remote Sensing Department, TU Berlin which was mainly focused on high accuracy registration between two images.