Skip to Content
SensorsSensors
  • Article
  • Open Access

4 February 2021

18 Pages

3D Convolutional Neural Networks Initialized from Pretrained 2D Convolutional Neural Networks for Classification of Industrial Parts

,
,
and
1
TECNALIA, Basque Research and Technology Alliance (BRTA), Mikeletegi Pasealekua 7, 20009 Donostia-San Sebastián, Spain
2
Robotics and Autonomous Systems Group, Universidad del País Vasco/Euskal Herriko Unibertsitatea, 48940 Basque, Spain
*
Author to whom correspondence should be addressed.

Abstract

Deep learning methods have been successfully applied to image processing, mainly using 2D vision sensors. Recently, the rise of depth cameras and other similar 3D sensors has opened the field for new perception techniques. Nevertheless, 3D convolutional neural networks perform slightly worse than other 3D deep learning methods, and even worse than their 2D version. In this paper, we propose to improve 3D deep learning results by transferring the pretrained weights learned in 2D networks to their corresponding 3D version. Using an industrial object recognition context, we have analyzed different combinations of 3D convolutional networks (VGG16, ResNet, Inception ResNet, and EfficientNet), comparing the recognition accuracy. The highest accuracy is obtained with EfficientNetB0 using extrusion with an accuracy of 0.9217, which gives comparable results to state-of-the art methods. We also observed that the transfer approach enabled to improve the accuracy of the Inception ResNet 3D version up to 18% with respect to the score of the 3D approach alone.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.