Source author record

Vikas Gupta

Vikas Gupta appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Computational Engineering, Finance, and Science Computer Vision Machine Learning math.NA Robotics

Catalog footprint

What is connected

3works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2020arXiv

Back to the Future: Joint Aware Temporal Deep Learning 3D Human Pose Estimation

We propose a new deep learning network that introduces a deeper CNN channel filter and constraints as losses to reduce joint position and motion errors for 3D video human body pose estimation. Our model outperforms the previous best result from the literature based on mean per-joint position error, velocity error, and acceleration errors on the Human 3.6M benchmark corresponding to a new state-of-the-art mean error reduction in all protocols and motion metrics. Mean per joint error is reduced by 1%, velocity error by 7% and acceleration by 13% compared to the best results from the literature. Our contribution increasing positional accuracy and motion smoothness in video can be integrated with future end to end networks without increasing network complexity. Our model and code are available at https://vnmr.github.io/ Keywords: 3D, human, image, pose, action, detection, object, video, visual, supervised, joint, kinematic

preprint2015arXiv

Local principle satisfying high order total variation diminishing approximation for non-sonic data extrema

The main contribution of this work is to construct higher than second order accurate total variation diminishing (TVD) schemes which can preserve high accuracy at non-sonic extrema with out induced local oscillations. It is done in the framework of local maximum principle (LMP) and non-conservative formulation. The representative uniformly second order accurate schemes are converted in to their non-conservative form using the ratio of consecutive gradient. These resulting schemes are analyzed for their non-linear LMP/TVD stability bounds using the local maximum principle. Based on the bounds, second order accurate hybrid numerical schemes are constructed using a shock detector. Numerical results are presented to show that such hybrid schemes yield TVD approximation with second or higher order convergence rate for smooth solution with extrema.

preprint2010arXiv

Computer Aided Design Modeling for Heterogeneous Objects

Heterogeneous object design is an active research area in recent years. The conventional CAD modeling approaches only provide geometry and topology of the object, but do not contain any information with regard to the materials of the object and so can not be used for the fabrication of heterogeneous objects (HO) through rapid prototyping. Current research focuses on computer-aided design issues in heterogeneous object design. A new CAD modeling approach is proposed to integrate the material information into geometric regions thus model the material distributions in the heterogeneous object. The gradient references are used to represent the complex geometry heterogeneous objects which have simultaneous geometry intricacies and accurate material distributions. The gradient references helps in flexible manipulability and control to heterogeneous objects, which guarantees the local control over gradient regions of developed heterogeneous objects. A systematic approach on data flow, processing, computer visualization, and slicing of heterogeneous objects for rapid prototyping is also presented.