Deep learning classification of large-scale point clouds: A case study on cuneiform tablets

Abstract

This paper introduces a novel network architecture for the classification of large-scale point clouds. The network is used to classify metadata from cuneiform tablets. As more than half a million tablets remain unprocessed, this can help create an overview of the tablets. The network is tested on a comparison dataset and obtains state-of-the-art performance. We also introduce new metadata classification tasks on which the network shows promising results. Finally, we introduce the novel Maximum Attention visualization, demonstrating that the trained network focuses on the intended features.

In-Hand Pose Estimation and Pin Inspection for Insertion of Through-Hole Components

“…To include details of the small pins, very dense and thereby large point clouds of the object are needed. The pin processing is, therefore, performed using a network that has shown good performance for very large point clouds [23]. The network has also shown an ability to focus on small details, which fits well with the small pins in our application.…”

“…The network has also shown an ability to focus on small details, which fits well with the small pins in our application. As the network in [23] was created for classification, the segmentation layer from PointNet [24] is added to enable instance segmentation. The training parameters are identical to the original paper [23], without the use of error weighing.…”

“…As the network in [23] was created for classification, the segmentation layer from PointNet [24] is added to enable instance segmentation. The training parameters are identical to the original paper [23], without the use of error weighing.…”