Fast Interactive Object Annotation with Curve-GCN
This addresses the problem of time-consuming object annotation for computer vision researchers and practitioners, offering a more efficient interactive tool.
The paper tackles the laborious task of manual object boundary annotation by proposing Curve-GCN, a framework that predicts all vertices simultaneously using a Graph Convolutional Network, supporting polygons or splines. It outperforms existing approaches in automatic mode and is significantly more efficient in interactive mode, running at 29.3ms (automatic) and 2.6ms (interactive), making it 10x and 100x faster than Polygon-RNN++.
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++.