知识卡片
手工与学习特征按数据条件选
学习特征可通过任务数据优化,手工几何或局部特征在训练资料不足、要求可解释或特定结构明显时仍有用途。网络不保证逐层显式复刻阈值、滤波与SIFT,也不是所有数据上必然优于手工方法;需要在目标场景比较。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 特征提取
原书出处 · 1 处
- · 查看原书依据
This chapter has introduced a variety of algorithms that turn high-dimensional input data into low-dimensional features, which can then be used to further reason about a problem. Recent advances in artificial neural networks (chapter 10) not only allow us to automatically train such feature detectors from data but also often outperform hand-coded feature detectors such as SIFT. Whereas the last decades have been dominated by
参考来源
- 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第09章《特征提取》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_9.xhtml)
- 结论依据:摘录中关于手工与学习特征按数据条件选的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:This chapter has introduced a variety of algorithms that turn high-dimensional input data into low-dimensional features, which can then be used to further reason about a problem. Recent advances in artificial neural networks (chapter 10) not only allow us to automatically train such feature detectors from data but also often outperform hand-coded feature detectors such as SIFT. Whereas the last decades have been dominated by