Comparison of Visual Datasets for Machine Learning
Comparison of Visual Datasets for Machine Learning
复制标题
机器学习视觉数据集的比较
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
Shu‐Ching Chen
中科院分区：
文献类型：
--
作者：
Kent W. Gauen;Ryan Dailey;John Laiman;Yuxiang Zi;Nirmal Asokan;Yung;G. Thiruvathukal;Mei;Shu‐Ching Chen
近年来最大的技术进步之一是使用机器学习处理视觉数据的快速进展。在促成这一发展的所有因素中，带有标签的数据集起着至关重要的作用。一些数据集被广泛地重复使用，用于研究和分析机器学习中的不同解决方案。许多系统，如自动驾驶汽车，依赖于使用机器学习来识别物体的组件。本文比较了机器学习的不同视觉数据集和框架。比较是定性和定量的，并调查对象检测标签的大小，位置和上下文信息。本文还提出了一种新的方法，使用实时的，地理标记的视觉数据创建数据集，大大提高了数据的上下文信息。这些数据可以通过交叉引用来自其他来源的信息（如天气）来自动标记。
One of the greatest technological improvements in recent years is the rapid progress using machine learning for processing visual data. Among all factors that contribute to this development, datasets with labels play crucial roles. Several datasets are widely reused for investigating and analyzing different solutions in machine learning. Many systems, such as autonomous vehicles, rely on components using machine learning for recognizing objects. This paper compares different visual datasets and frameworks for machine learning. The comparison is both qualitative and quantitative and investigates object detection labels with respect to size, location, and contextual information. This paper also presents a new approach creating datasets using real-time, geo-tagged visual data, greatly improving the contextual information of the data. The data could be automatically labeled by cross-referencing information from other sources (such as weather).