Spiking Neural Network Based Object Pose Alignment

Sushant Yadav, Chandarjeet Singh Chundawat, Santosh Chaudhary, Rajesh Kumar

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Object pose alignment, a prerequisite for many computer vision tasks, e.g., face recognition, 3D face reconstruction, robotics, augmented reality, etc. There are lot of research to address this issue, still, there are still numerous issues regarding the problem. Among which one of them is the computational efficiency. To address this issue, this article proposes a novel method for object pose alignment of 6 DoF with Spiking Neural Network (SNN). SNNs are biologically inspired neural networks which replace traditional networks through their energy efficiency and event-driven processing mechanism. The method uses SNN to predict the translation and rotation coordinates for the base position to align with the ground truth pose. The proposed method shows potential by reducing the computational cost by almost 10% and the final results of the problem are represented in the result section, representing the initial pose and the aligned pose after training with SNN.

Original languageEnglish
Title of host publicationSoft Computing
Subtitle of host publicationTheories and Applications - Proceedings of SoCTA 2024
EditorsRajesh Kumar, Ajit Kumar Verma, Om Prakash Verma, Jitendra Rajpurohit
PublisherSpringer Science and Business Media Deutschland GmbH
Pages389-397
Number of pages9
ISBN (Print)9789819659548
DOIs
Publication statusPublished - 2025
Event9th International Conference on Soft Computing: Theories and Applications, SoCTA 2024 - Jaipur, India
Duration: 27 Dec 202429 Dec 2024

Publication series

NameLecture Notes in Networks and Systems
Volume1343 LNNS
ISSN (Print)2367-3370
ISSN (Electronic)2367-3389

Conference

Conference9th International Conference on Soft Computing: Theories and Applications, SoCTA 2024
Country/TerritoryIndia
CityJaipur
Period27/12/2429/12/24

Keywords

  • 3D computer vision
  • Object pose alignment
  • Spiking neural network

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Signal Processing
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Spiking Neural Network Based Object Pose Alignment'. Together they form a unique fingerprint.

Cite this