A Method of Detecting Interaction Between Breeding cows from Time-series Bird’s Eye Pictures of Pasture

A Method of Detecting Interaction Between Breeding cows from Time-series Bird’s Eye Pictures of Pasture
复制标题

一种从牧场时间序列鸟瞰图检测种牛间相互作用的方法

DOI:
10.1109/ijcnn.2016.7727576
复制
发表时间:
2016
期刊:
Proceedings of the 2016 International Joint Conference on Neural Networks
影响因子:
--
通讯作者:
and Takenao Ohkawa
and Takenao Ohkawa
中科院分区:
--
文献类型:
--
作者:
Yoji Hosomi;Kenji Oyama;Takeshi Honda;and Takenao Ohkawa

文献摘要

相似文献

众所周知,种牛具有社会性,它们相互作用并形成群体。本文提出了一种方法,检测繁殖奶牛之间的相互作用,从时间序列的牧场图片的相似图像检索方法,使用视觉词袋。我们将交互检测分为三个任务:检测交互中的一对奶牛,精确定位观察到交互的时间和地点以及识别其类型。本文主要研究识别任务。在所提出的方法中,一个类似的图像检索方法被用来分类的互动,和信息的前,后的互动被认为是提高分类的准确性。实验证明了该方法的有效性。
Breeding cows are known to engage in sociality, in which they interact and form groups. This paper proposes a method of detecting the interaction between breeding cows from time-series pictures of pastures by a similar image retrieval method using a Bag of Visual Words. We divided the interaction detection into three tasks: detecting a pair of cows in an interaction, pinpointing the time and the place that the interaction was observed and identifying its type. This paper focuses on the identifying task. In the proposed method, a similar image retrieval method is used to classify the interaction, and the information from the states of preceding and succeeding the interaction is considered to improve the accuracy of the interaction classification. The effectiveness of the proposed method was confirmed experimentally.