Dense Extreme Inception Network for Edge Detection

4 Dec 2021  ·  Xavier Soria, Angel Sappa, Patricio Humanante, Arash Akbarinia ·

<<<This is a pre-acceptance version, please, go through Pattern Recognition Journal on Sciencedirect to read the final version>>>. Edge detection is the basis of many computer vision applications. State of the art predominantly relies on deep learning with two decisive factors: dataset content and network's architecture. Most of the publicly available datasets are not curated for edge detection tasks. Here, we offer a solution to this constraint. First, we argue that edges, contours and boundaries, despite their overlaps, are three distinct visual features requiring separate benchmark datasets. To this end, we present a new dataset of edges. Second, we propose a novel architecture, termed Dense Extreme Inception Network for Edge Detection (DexiNed), that can be trained from scratch without any pre-trained weights. DexiNed outperforms other algorithms in the presented dataset. It also generalizes well to other datasets without any fine-tuning. The higher quality of DexiNed is also perceptually evident thanks to the sharper and finer edges it outputs.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Edge Detection BIPED DexiNed ODS 0.895 # 1
Number of parameters (M) 35M # 5
Edge Detection MDBD DexiNed-a ODS 0.894 # 1
Edge Detection MDBD DexiNed-f ODS 0.891 # 2
Edge Detection UDED DexiNed ODS 0.815 # 3

Methods


No methods listed for this paper. Add relevant methods here