Model pruning in depth completion CNNs for forestry robotics with simulated annealing

Andrada, M.E., Ferreira, J.F. ORCID: 0000-0002-2510-2412, Kantor, G., Portugal, D. and Antunes, C.H., 2022. Model pruning in depth completion CNNs for forestry robotics with simulated annealing. In: Innovation in Forestry Robotics: Research and Industry Adoption Workshop - IEEE Conference on Robotics and Automation (ICRA 2022), Philadelphia (PA), USA, 23-27 May 2022.

[img]
Preview
Text
1551539_Ferreira.pdf - Post-print

Download (3MB) | Preview

Abstract

In this article, we present an analysis of model compression in depth completion neural networks for forestry robotics, considering the increasing demands of real time autonomous solutions. Specifically, we implement a single state simulated annealing meta-heuristic for model pruning in the ENet and MSG-CHN neural networks for depth completion. We run experiments in three different datasets and analyze how different levels of pruning affect the accuracy and speed of the models. Experimental tests show that increasing sparsity has different effects depending on the neural network and dataset. ENet has negligible difference in accuracy and it would greatly benefit from lowering the amount of FLOPs, while MSG-CHN displays an inconsistent behavior depending on the dataset. This suggests that while both models benefit from model compression techniques, the optimal sparsity level depends on environment, dataset and neural network.

Item Type: Conference contribution
Creators: Andrada, M.E., Ferreira, J.F., Kantor, G., Portugal, D. and Antunes, C.H.
Date: May 2022
Identifiers:
NumberType
1551539Other
Divisions: Schools > School of Science and Technology
Record created by: Laura Ward
Date Added: 08 Jun 2022 07:58
Last Modified: 08 Jun 2022 07:58
Related URLs:
URI: http://irep.ntu.ac.uk/id/eprint/46419

Actions (login required)

Edit View Edit View

Views

Views per month over past year

Downloads

Downloads per month over past year