arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.05701cs.CVcs.LGeess.IV
收藏

RetinotopicNet:一种基于局部描述符和全局上下文的迭代注意力机制

RetinotopicNet: An Iterative Attention Mechanism Using Local Descriptors with Global Context

Aptiv(安波福);Mercator School of Management, University Duisburg-Essen(杜伊斯堡-埃森大学墨卡托管理学院)

Thomas Kurbiel, Shahrzad Khaleghian

摘要:

卷积神经网络(CNN)是近年来计算机视觉研究中许多进步背后的驱动力。这一进展催生了许多实际应用,我们看到越来越需要将CNN有效地迁移到当今的嵌入式系统中。然而,传统的CNN缺乏尺度和旋转不变性的特性:这是自然图像中最常见的两种变换。因此,CNN必须在不同的尺度上学习相同对象的不同特征。这种冗余是CNN需要非常深以达到所需精度的主要原因。在本文中,我们开发了一个有效的解决方案,通过复制自然如何解决人脑中的问题。为此,我们让我们的CNN对使用对数极坐标变换提取的小块进行操作,这是已知的尺度和旋转等变。这种方法提取的图像块具有中心区域放大、边缘区域压缩的良好特性。因此,我们获得具有全局上下文信息的局部描述符。然而,单个块的处理通常不足以在例如分类任务中实现高精度。因此,我们连续跳到几个不同的位置,称为扫视,从而建立对整个图像的理解。由于对数极坐标块包含全局上下文信息,因此我们可以仅使用小块有效地计算后续扫视。扫视有效地补偿了对数极坐标变换的平移等方差的缺乏。

英文摘要:

Convolutional Neural Networks (CNNs) were the driving force behind many advancements in Computer Vision research in recent years. This progress has spawned many practical applications and we see an increased need to efficiently move CNNs to embedded systems today. However traditional CNNs lack the property of scale and rotation invariance: two of the most frequently encountered transformations in natural images. As a consequence CNNs have to learn different features for same objects at different scales. This redundancy is the main reason why CNNs need to be very deep in order to achieve the desired accuracy. In this paper we develop an efficient solution by reproducing how nature has solved the problem in the human brain. To this end we let our CNN operate on small patches extracted using the log-polar transform, which is known to be scale and rotation equivariant. Patches extracted in this way have the nice property of magnifying the central field and compressing the periphery. Hence we obtain local descriptors with global context information. However the processing of a single patch is usually not sufficient to achieve high accuracies in e.g. classification tasks. We therefore successively jump to several different locations, called saccades, thus building an understanding of the whole image. Since log-polar patches contain global context information, we can efficiently calculate following saccades using only the small patches. Saccades efficiently compensate for the lack of translation equivariance of the log-polar transform.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑