Prepare

• Win10 1809 17763.973
• VirtualBox 6.0.8
• android-x86-9.0-rc2.iso

Note: I really recommend you to download from official website.

Note: This article can only make android-x86 support armeabi, not armeabi-v7a or armeabi-v8a

Section 1.1

• 在图像中直线保持原来的特征，而诸如形状、角度、距离、比例则会发生改变

Part 0 Outline

• Chapter 2 introduces projective transformations of 2-space.
• Chapter 3 covers the projective geometry of 3-space.
• Chapter 4 introducesestimationof geometry from image m easureme
• Chapter 5 describes how the results of estimation algorithms may be evaluated

link

Reidentification by Relative Distance Comparison

Challenge:

• large visual appearance changes caused by variations in view angle, lighting, background clutter, and occlusion
• 之前的大部分算法寻找独特的视觉特征。但寻找在数据规模大、现实条件不同的数据集中能够保持鲁棒性的视觉特征仍然十分困难。
• 在不同条件下，有些特征比其他特征更重要，更稳定，使用l1-Norm等普遍采用的标准的距离评估方法并不合适，因为它们会等权重地对待所有特征。

In order to find a correc match Given a query image of a person:

• First, a feature representation is computed from both the query and each of the gallery images.
• Second, the distance between each pair of potential matches is measured

源地址

arXiv:1712.07465: Recurrent Attentional Reinforcement Learning for Multi-label Image Recognition

简介

识别图像中的多个标签是计算机视觉中的一项基本但具有挑战性的任务。针对现有方法计算成本高、不能有效利用空间上下文的问题，论文提出了循环迭代的结合注意力机制的强化学习框架，并进行了对应的熔断测试。