Semantic Edge Detection with Diverse Deep Supervision
This work addresses semantic edge detection for applications like semantic segmentation and object recognition, representing an incremental improvement over existing methods.
The paper tackles the problem of semantic edge detection by proposing a fully convolutional neural network with diverse deep supervision to address the challenge of simultaneously locating fine edges and identifying high-level semantics, achieving improved results on SBD and Cityscapes datasets.
Semantic edge detection (SED), which aims at jointly extracting edges as well as their category information, has far-reaching applications in domains such as semantic segmentation, object proposal generation, and object recognition. SED naturally requires achieving two distinct supervision targets: locating fine detailed edges and identifying high-level semantics. Our motivation comes from the hypothesis that such distinct targets prevent state-of-the-art SED methods from effectively using deep supervision to improve results. To this end, we propose a novel fully convolutional neural network using diverse deep supervision (DDS) within a multi-task framework where bottom layers aim at generating category-agnostic edges, while top layers are responsible for the detection of category-aware semantic edges. To overcome the hypothesized supervision challenge, a novel information converter unit is introduced, whose effectiveness has been extensively evaluated on SBD and Cityscapes datasets.