A Static Video Summarization Approach for the Analysis of Cattle's Movement

DOI Web Site

この論文をさがす

抄録

<p>Today, in image processing and pattern recognition, various research studies have been conducted, and many results have been reported. Video summarization technology is also a significant area and plays an essential role in processing vast amounts of video. This paper proposes a static video summarization approach for analysis and visualization of calf behavior. In cattle management, it is important to find the behavioral characteristics of individual cattle. However, many researchers have struggled because they require long-term observations and are a tedious and time-consuming task for manual observation, recording, and analyzing behavioral data.</p><p>The authors attempted to attach a camera to the calf halter, record a video of the calf's view, and extract various calf's behavior features by image processing of the video. This method uses no specialized hardware, and the captured video can be repeatedly analyzed offline with different ideas.</p><p>In this paper, a video summarization technique based on the κ-means clustering algorithm was used to analyze calf movements on an experimental barn from the video. The calf's places stayed for a long time, or the places where the calf visited frequently were detected as a cluster, and the calf's movement on the farm could be visualized as a map by the transition between clusters.</p>

収録刊行物

  • 日本画像学会誌

    日本画像学会誌 60 (1), 9-16, 2021-02-10

    一般社団法人 日本画像学会

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ