Object Contour Detection with a Fully Convolutional Encoder-Decoder Network

TL;DR

This paper introduces a fully convolutional encoder-decoder network for detecting higher-level object contours, achieving superior precision and generalization in object proposal generation compared to previous methods.

Contribution

The paper presents a novel deep learning model trained end-to-end for object contour detection, outperforming prior low-level edge detection techniques and enhancing object proposal quality.

Findings

Higher precision in object contour detection than previous methods

Good generalization to unseen object classes within the same super-categories

Significant improvement in average recall for object proposals on PASCAL VOC

Abstract

We develop a deep learning algorithm for contour detection with a fully convolutional encoder-decoder network. Different from previous low-level edge detection, our algorithm focuses on detecting higher-level object contours. Our network is trained end-to-end on PASCAL VOC with refined ground truth from inaccurate polygon annotations, yielding much higher precision in object contour detection than previous methods. We find that the learned model generalizes well to unseen object classes from the same super-categories on MS COCO and can match state-of-the-art edge detection on BSDS500 with fine-tuning. By combining with the multiscale combinatorial grouping algorithm, our method can generate high-quality segmented object proposals, which significantly advance the state-of-the-art on PASCAL VOC (improving average recall from 0.62 to 0.67) with a relatively small amount of candidates ($\sim$1660 per image).