Semantic Segmentation from Limited Training Data
Authors:
- 1 May 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1908-1915
Abstract
We present our approach for robotic perception in cluttered scenes that led to winning the recent Amazon Robotics Challenge (ARC) 2017. Next to small objects with shiny and transparent surfaces, the biggest challenge of the 2017 competition was the introduction of unseen categories. In contrast to traditional approaches which require large collections of annotated data and many hours of training, the task here was to obtain a robust perception pipeline with only few minutes of data acquisition and training time. To that end, we present two strategies that we explored. One is a deep metric learning approach that works in three separate steps: semantic-agnostic boundary detection, patch classification and pixel-wise voting. The other is a fully-supervised semantic segmentation approach with efficient dataset collection. We conduct an extensive analysis of the two methods on our ARC 2017 dataset. Interestingly, only few examples of each class are sufficient to fine-tune even very deep convolutional neural networks for this specific task.Keywords
All Related Versions
This publication has 27 references indexed in Scilit:
- Analysis and Observations From the First Amazon Picking ChallengeIEEE Transactions on Automation Science and Engineering, 2016
- Lessons from the Amazon Picking Challenge: Four Aspects of Building Robotic Systems2016
- Weakly-and Semi-Supervised Learning of a Deep Convolutional Network for Semantic Image Segmentation2015
- Holistically-Nested Edge Detection2015
- Learning descriptors for object recognition and 3D pose estimation2015
- From image-level to pixel-level labeling with Convolutional Networks2015
- Going deeper with convolutions2015
- Fully convolutional networks for semantic segmentation2015
- Indoor Segmentation and Support Inference from RGBD Images2012
- Boundary Extraction in Natural Images Using Ultrametric Contour Maps2006