Fast Interactive Object Annotation With Curve-GCN

Huan Ling, Jun Gao, Amlan Kar, Wenzheng Chen, Sanja Fidler; The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 5257-5266


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, and runs in real time. 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 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++.

Related Material

author = {Ling, Huan and Gao, Jun and Kar, Amlan and Chen, Wenzheng and Fidler, Sanja},
title = {Fast Interactive Object Annotation With Curve-GCN},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}