Object Bank: A High-Level Image Representation for Scene Classification & Semantic Feature Sparsification
Object Bank: A High-Level Image Representation for Scene Classification & Semantic Feature Sparsification
复制标题
DOI：
--
复制
发表时间：
2010-12
期刊：
影响因子：
--
通讯作者：
Li-Jia Li-Li-Jia-Li-2040091191;Hao Su;E. Xing;Li Fei-Fei-Li-Fei-Fei-48004138
中科院分区：
文献类型：
--
作者：
Li-Jia Li-Li-Jia-Li-2040091191;Hao Su;E. Xing;Li Fei-Fei-Li-Fei-Fei-48004138
鲁棒的低级别图像特征已被证明是各种视觉识别任务（如对象识别和场景分类）的有效表示;但像素，甚至局部图像补丁，几乎没有语义意义。对于高级视觉任务，这种低级图像表示可能是不够的。在本文中，我们提出了一种高级图像表示，称为对象库，其中图像表示为大量预先训练的通用对象检测器的尺度不变响应图，对测试数据集或视觉任务是盲目的。利用对象库表示，高级视觉识别任务的上级性能可以通过简单的现成分类器（例如逻辑回归和线性SVM）来实现。稀疏算法使我们的表示更有效和可扩展的大型场景数据集，并揭示语义有意义的特征模式。
Robust low-level image features have been proven to be effective representations for a variety of visual recognition tasks such as object recognition and scene classification; but pixels, or even local image patches, carry little semantic meanings. For high level visual tasks, such low-level image representations are potentially not enough. In this paper, we propose a high-level image representation, called the Object Bank, where an image is represented as a scale-invariant response map of a large number of pre-trained generic object detectors, blind to the testing dataset or visual task. Leveraging on the Object Bank representation, superior performances on high level visual recognition tasks can be achieved with simple off-the-shelf classifiers such as logistic regression and linear SVM. Sparsity algorithms make our representation more efficient and scalable for large scene datasets, and reveal semantically meaningful feature patterns.