ContourRend: A Segmentation Method for Improving Contours by Rendering
This work addresses the issue of poor contour quality in segmentation for applications like autonomous driving, but it is incremental as it builds on existing GCN-based methods with a specific refinement technique.
The paper tackles the problem of blurry edges in object segmentation by proposing ContourRend, a method that uses a contour renderer to refine segmentation contours, achieving a mean IoU of 72.41% on the Cityscapes dataset, which is a 1.22% improvement over the baseline.
A good object segmentation should contain clear contours and complete regions. However, mask-based segmentation can not handle contour features well on a coarse prediction grid, thus causing problems of blurry edges. While contour-based segmentation provides contours directly, but misses contours' details. In order to obtain fine contours, we propose a segmentation method named ContourRend which adopts a contour renderer to refine segmentation contours. And we implement our method on a segmentation model based on graph convolutional network (GCN). For the single object segmentation task on cityscapes dataset, the GCN-based segmentation con-tour is used to generate a contour of a single object, then our contour renderer focuses on the pixels around the contour and predicts the category at high resolution. By rendering the contour result, our method reaches 72.41% mean intersection over union (IoU) and surpasses baseline Polygon-GCN by 1.22%.