Cognitive Contour Detection of Sparse-Structured Objects in the Alpha-Shape Scale Space

This paper proposes a novel scale-space method for detecting cognitive contours of sparse-structured objects. Additionally, a benchmark image dataset of sparse-structured objects is constructed in this paper, which is publicly available for evaluating cognitive contour detection performance. Here are the code and the dataset for this paper.

Dateset

Download the SOIS dataset from BaiDuNetdisk or Google Drive.

Requirements

• Python
  • scipy
  • numba
  • cv2
  • copy

How to use

• This demo can be easily used by running the 'demo.py' file.

python demo.py

• We will release the details once the paper is accepted.