BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentation

TL;DR

BoxSup is a novel method that uses bounding box annotations instead of pixel-level masks to train convolutional networks for semantic segmentation, achieving competitive and state-of-the-art results.

Contribution

We introduce BoxSup, a technique that iteratively generates segmentation masks from bounding boxes to train deep networks without requiring pixel-level annotations.

Findings

Achieves results comparable to fully supervised methods using only bounding boxes.

Outperforms previous weakly supervised approaches on PASCAL datasets.

Enables leveraging large amounts of bounding box data for improved segmentation.

Abstract

Recent leading approaches to semantic segmentation rely on deep convolutional networks trained with human-annotated, pixel-level segmentation masks. Such pixel-accurate supervision demands expensive labeling effort and limits the performance of deep networks that usually benefit from more training data. In this paper, we propose a method that achieves competitive accuracy but only requires easily obtained bounding box annotations. The basic idea is to iterate between automatically generating region proposals and training convolutional networks. These two steps gradually recover segmentation masks for improving the networks, and vise versa. Our method, called BoxSup, produces competitive results supervised by boxes only, on par with strong baselines fully supervised by masks under the same setting. By leveraging a large amount of bounding boxes, BoxSup further unleashes the power of deep convolutional networks and yields state-of-the-art results on PASCAL VOC 2012 and PASCAL-CONTEXT.