seavan的专栏

Auto-context and Its Application to High-level Vision Tasks

SLIC : SLIC Superpixels Compared to State-of-the-art Superpixel Methods

(JAIR 2002) SMOTE Synthetic Minority Over-sampling Technique这篇文章讨论了样本不均衡问题。思路很简单清晰。简单介绍一下：样本不均衡是指在分类过程中，有些类拥有大量的训练样本，有些类则只有少数的样本，让分类器学习起来很困难。而这样的情况又是最常见的，比如说二分类：要对图像中检测到的物体识别是苹果还是非苹果，那么非类作为负样本具有无穷

Neighbourhood Components Analysis

(CVPR 2015) DeepContour: A Deep Convolutional Feature Learned by Positive-sharing Loss for Contour Detection

Sketch Tokens: A Learned Mid-level Representation for Contour and Object Detection

reading list... 加油读论文