Skip to content

code of "Saliency Detection via CNN Coarse Learning and Compactness Based ELM Refinement"

Notifications You must be signed in to change notification settings

DIYer22/salience-ELM

Repository files navigation

使用超像素+紧密度算子+传统分类器的图像显著区域检测

环境: Python 2.7

依赖: pip install boxx hpelm;conda install opencv

测试: python test.py

Update in 2018.12

调用命令:

cd ../salience-ELM
source activate /home/yanglei/miniconda2
rlaunch -P 50 --cpu=1 --memory=15240  -- python algorithm.py
source deactivate

Update in 2018.08

注意:

  1. from boxx import * 可能会和早期版本冲突
  2. algorithm.buildMethodDic 是个函数字典, 存储各个方法, 比较重要
  3. coarseMethods=[] 则为 Compactness 原文的原始方法

Old Version (2016)

基于ELM的图像显著区域检测


作者:小磊

邮箱:ylxx@live.com

时间:2016-11-20

环境

Python版本:Python 2.7 with Ipython

常见库:numpy,skiamge

ELM库: hpelm

下载地址:https://pypi.python.org/pypi/hpelm(建议手动安装)

文档:http://hpelm.readthedocs.io/en/latest/api/elm.html

数据集:http://202.118.75.4/lu/DUT-OMRON/index.htm

注意事项

将数据集的DUT-OMRON-imagepixelwiseGT-new-PNG文件夹放到 项目父文件夹下

(参考调用代码imgDir = '../DUT-OMRON-image/')

About

code of "Saliency Detection via CNN Coarse Learning and Compactness Based ELM Refinement"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages