Manually labeling objects by tracing their boundaries is a laborious process. In Polygon-RNN++ the authors proposed Polygon-RNN that produces polygonal annotations in a recurrent manner using a CNN-RNN architecture, allowing interactive correction via humans-in-the-loop. We propose a new framework that alleviates the sequential nature of Polygon-RNN, by predicting all vertices simultaneously using a Graph Convolutional Network (GCN). Our model is trained end-to-end. It supports object annotation by either polygons or splines, facilitating labeling efficiency for both line-based and curved objects. We show that Curve-GCN outperforms all existing approaches in automatic mode, including the powerful PSP-DeepLab and is significantly more efficient in interactive mode than Polygon-RNN++. Our model runs at 29.3ms in automatic, and 2.6ms in interactive mode, making it 10x and 100x faster than Polygon-RNN++.
@article{arxiv.1903.06874,
title = {Fast Interactive Object Annotation with Curve-GCN},
author = {Huan Ling and Jun Gao and Amlan Kar and Wenzheng Chen and Sanja Fidler},
journal= {arXiv preprint arXiv:1903.06874},
year = {2019}
}
Comments
In Computer Vision and Pattern Recognition (CVPR), Long Beach, US, 2019