Journal
INTELLIGENT DECISION TECHNOLOGIES, KES-IDT 2021
Volume 238, Issue -, Pages 209-219Publisher
SPRINGER INTERNATIONAL PUBLISHING AG
DOI: 10.1007/978-981-16-2765-1_17
Keywords
Printed circuit boards (PCBs); Convolutional neural network (CNN); Homogeneous vector capsules (HVCs); Capsule; Data augmentation
Ask authors/readers for more resources
The study presents a dataset of high-resolution images of 13 micro-PCBs, labeled for rotations and perspectives, and experiments show that training a neural network equipped with HVCs and diverse perspectives achieve the highest classification accuracy on micro-PCB data.
We present a dataset consisting of high-resolution images of 13 micro-PCBs captured in various rotations and perspectives relative to the camera, with each sample labeled for PCB type, rotation category, and perspective categories. We then present the design and results of experimentation on combinations of rotations and perspectives used during training and the resulting impact on test accuracy. We then show when and how well data augmentation techniques are capable of simulating rotations versus perspectives not present in the training data. We perform all experiments using CNNs with and without homogeneous vector capsules (HVCs) and investigate and show the capsules' ability to better encode the equivariance of the sub-components of the micro-PCBs. The results of our experiments lead us to conclude that training a neural network equipped with HVCs, capable of modeling equivariance among sub-components, coupled with training on a diversity of perspectives, achieves the greatest classification accuracy on micro-PCB data.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available