Showing 1–1 of 1 results for author: Lindqvist, L
-
Depth Contrast: Self-Supervised Pretraining on 3DPM Images for Mining Material Classification
Authors:
Prakash Chandra Chhipa,
Richa Upadhyay,
Rajkumar Saini,
Lars Lindqvist,
Richard Nordenskjold,
Seiichi Uchida,
Marcus Liwicki
Abstract:
This work presents a novel self-supervised representation learning method to learn efficient representations without labels on images from a 3DPM sensor (3-Dimensional Particle Measurement; estimates the particle size distribution of material) utilizing RGB images and depth maps of mining material on the conveyor belt. Human annotations for material categories on sensor-generated data are scarce a…
▽ More
This work presents a novel self-supervised representation learning method to learn efficient representations without labels on images from a 3DPM sensor (3-Dimensional Particle Measurement; estimates the particle size distribution of material) utilizing RGB images and depth maps of mining material on the conveyor belt. Human annotations for material categories on sensor-generated data are scarce and cost-intensive. Currently, representation learning without human annotations remains unexplored for mining materials and does not leverage on utilization of sensor-generated data. The proposed method, Depth Contrast, enables self-supervised learning of representations without labels on the 3DPM dataset by exploiting depth maps and inductive transfer. The proposed method outperforms material classification over ImageNet transfer learning performance in fully supervised learning settings and achieves an F1 score of 0.73. Further, The proposed method yields an F1 score of 0.65 with an 11% improvement over ImageNet transfer learning performance in a semi-supervised setting when only 20% of labels are used in fine-tuning. Finally, the Proposed method showcases improved performance generalization on linear evaluation. The implementation of proposed method is available on GitHub.
△ Less
Submitted 18 October, 2022;
originally announced October 2022.
