
- This event has passed.
{Seminar} @ CDS: 16th January: ” Robust Registration, Reconstruction and Reverse Engineering of 3D Shapes/Scene”
16 Jan @ 11:30 AM -- 12:30 PM
Department of Computational and Data Sciences
Department Seminar
SPEAKER : Dr. Sk Aziz Ali
TITLE : “Robust Registration, Reconstruction and Reverse Engineering of 3D Shapes/Scene.“
Date & Time : January 16, 2023, 11:30 AM.
Venue : #102, CDS Seminar Hall.
===========================================================================================
ABSTRACT
3D scene/shape reconstruction, understanding, and manipulation are of high importance in computer vision and graphics research. Fast and robust 3D shape reconstruction and registration are plenary methods that are widely used for obtaining expressive 3D model of a scene or object. This talk will present a novel learning representation of 3D shapes leveraging – (I) fast, robust and continuous alignment of scene fragments/partial scans of an object and (II) its dense reconstruction. Supervised learning of relative sensor pose is integrated for the frame-to-frame scan alignment. The network architecture is inspired by novel gravitational potential energy optimization to estimate rigid body dynamics between a source and a target shape. The same gravitational energy function is naturally extensible to capture isomorphic shape deformation and high-fidelity clothing deformation. For non-stop reconstruction of outdoor scenes, a life-long LiDAR-based SLAM method is
designed which has an in-built predictive uncertainty estimator and a general-purpose and lightweight correspondence matcher. In this context, the presentation will also cover how modern Artec3D sensors for
high-precision shape scanning help in generating datasets of diverse 3D body shapes and mechanical CAD models.
BIOGRAPHY
Sk Aziz Ali has studied Computer Graphics and Vision at the Technical University of Kaiserslautern (TU KL), Germany. He received his master’s degree in computer science in 2017 and completed the doctoral
dissertation in 2022 from the same University. He did his master thesis, titled as – “Visualization of FEM data for FRP materials”, from the German Aerospace (DLR) where he prototyped a visually based Fiber
Reinforced Plastic (FRP) material failure analysis tool to help structure designing engineers analyzing the strength of aircraft components. Sk Aziz Ali joined the Augmented Vision research group at the German Research Center for Artificial Intelligence (DFKI) in 2017 during his PhD. He is currently holding Research Associate position at the Computer Vision, Imaging and Machine Intelligence (CVI2) research group in SnT, University of Luxembourg. His primary research interest combines – Registration and Reconstruction of 3D Shapes/Scene, LiDAR-based SLAM, Deep Learning-Based Reverse CAD modeling.
Host Faculty: Prof. Venkatesh Babu
==================================================================================
ALL ARE WELCOME