Skip to content
/ BSCP Public

Shape recognition based on the combination of contour and skeleton (codes for our paper on PRL2016)

Notifications You must be signed in to change notification settings

jy9387/BSCP

Repository files navigation

BSCP

Shape recognition based on the combination of contour information and skeleton information.

Codes for our PRL2016 paper: "Shape Recognition by Bag of Skeleton-associated Contour Parts". [paper]

bscp_pipeline

Requirements:

  1. Dataset: we use Animal Dataset [paper] for evaluating. The dataset contains 2000 images, which are divided into 20 classes. We extend the dataset by adding the corresponding skeleton-associated contour maps. Please [download] it and extract it to the "data/" directory for training and testing.

  2. Tools: you might download the pdollar's toolbox [download] and the liblinear [download], and put them into the "include/" directory.

  3. Interface: please see exp_animal.m

For more technical details please refer to our paper.

If you find bugs or have questions, please contact me. Glad to hear from you.


Mail: [email protected]

Homepage: https://jy9387.github.io/

About

Shape recognition based on the combination of contour and skeleton (codes for our paper on PRL2016)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published