Evaluating the performance of semiautomated photographic identification programs for leopard seals

IF 1.5 4区环境科学与生态学 Q3 Environmental Science

Wildlife Society Bulletin Pub Date : 2024-03-27 DOI:10.1002/wsb.1520

Alexander A. Grabham, Krista van der Linde, Ximena J. Nelson

{"title":"Evaluating the performance of semiautomated photographic identification programs for leopard seals","authors":"Alexander A. Grabham, Krista van der Linde, Ximena J. Nelson","doi":"10.1002/wsb.1520","DOIUrl":null,"url":null,"abstract":"The photographic identification (photo-ID) of individual animals can be time-consuming and erroneous. Recent efforts to collect photographs of leopard seals (Hydrurga leptonyx) from across their range have necessitated the technological streamlining of photo-ID. We constructed a dataset containing 595 photographs of the head and body of individuals recorded in New Zealand between 2008 and 2022 to test the performance of semiautomated 3 photo-ID programs: HotSpotter, Interactive Individual Identification System's Pattern+ (I3S), and Wild-ID. We classified attributes of photographs (e.g., quality) and individuals (i.e., pelage patterns) to assess their effect on performance. We compared performance using Top20 and Top1 Accuracy, defined as the proportion of test photographs where the highest ranked correct identity was in the top 20 and top 1, respectively, matched reference photographs. HotSpotter outperformed I3S and Wild-ID in both Top20 and Top1 Accuracy of most assessed attributes. Maximizing HotSpotter's performance may be achieved through several methods, including increasing the number and variety of photographs of individuals in the dataset. HotSpotter will likely perform better with photographs without obstructions (e.g., debris from beaches), such as on the pack ice of their primary Antarctic range. We highlight the viability of HotSpotter in assisting the photo-ID of leopard seals, and more broadly, other species with similar markings.","PeriodicalId":23845,"journal":{"name":"Wildlife Society Bulletin","volume":"45 1","pages":""},"PeriodicalIF":1.5000,"publicationDate":"2024-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Wildlife Society Bulletin","FirstCategoryId":"93","ListUrlMain":"https://doi.org/10.1002/wsb.1520","RegionNum":4,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"Environmental Science","Score":null,"Total":0}

引用次数: 0

Abstract

The photographic identification (photo-ID) of individual animals can be time-consuming and erroneous. Recent efforts to collect photographs of leopard seals (Hydrurga leptonyx) from across their range have necessitated the technological streamlining of photo-ID. We constructed a dataset containing 595 photographs of the head and body of individuals recorded in New Zealand between 2008 and 2022 to test the performance of semiautomated 3 photo-ID programs: HotSpotter, Interactive Individual Identification System's Pattern⁺ (I³S), and Wild-ID. We classified attributes of photographs (e.g., quality) and individuals (i.e., pelage patterns) to assess their effect on performance. We compared performance using Top20 and Top1 Accuracy, defined as the proportion of test photographs where the highest ranked correct identity was in the top 20 and top 1, respectively, matched reference photographs. HotSpotter outperformed I³S and Wild-ID in both Top20 and Top1 Accuracy of most assessed attributes. Maximizing HotSpotter's performance may be achieved through several methods, including increasing the number and variety of photographs of individuals in the dataset. HotSpotter will likely perform better with photographs without obstructions (e.g., debris from beaches), such as on the pack ice of their primary Antarctic range. We highlight the viability of HotSpotter in assisting the photo-ID of leopard seals, and more broadly, other species with similar markings.

Abstract Image

查看原文本刊更多论文

评估豹纹海豹半自动摄影识别程序的性能

对动物个体进行照片识别（photo-ID）既费时又容易出错。最近，为了收集豹纹海豹（Hydrurga leptonyx）在其分布区的照片，有必要对照片识别进行技术简化。我们建立了一个数据集，其中包含 2008 年至 2022 年期间在新西兰记录的 595 张豹斑海豹个体头部和身体的照片，用于测试 3 种照片识别程序的半自动化性能：HotSpotter、交互式个体识别系统 Pattern+ (I3S) 和 Wild-ID。我们对照片的属性（如质量）和个体的属性（如皮层模式）进行了分类，以评估它们对性能的影响。我们使用 Top20 和 Top1 准确率对性能进行了比较，Top20 和 Top1 准确率的定义是：在与参考照片匹配的测试照片中，最高正确率排名分别在前 20 位和前 1 位的比例。在大多数评估属性的 Top20 和 Top1 准确率方面，HotSpotter 的表现均优于 I3S 和 Wild-ID。最大限度地提高 HotSpotter 的性能可以通过多种方法实现，包括增加数据集中个体照片的数量和种类。HotSpotter在拍摄无遮挡物（如海滩上的碎石）的照片时可能会有更好的表现，例如在其主要南极分布区的冰包上。我们强调 HotSpotter 在协助豹纹海豹以及更广泛意义上的具有类似标记的其他物种的照片识别方面的可行性。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Wildlife Society Bulletin BIODIVERSITY CONSERVATION-

CiteScore

2.10

自引率

13.30%

发文量

期刊介绍： The Wildlife Society Bulletin is a journal for wildlife practitioners that effectively integrates cutting edge science with management and conservation, and also covers important policy issues, particularly those that focus on the integration of science and policy. Wildlife Society Bulletin includes articles on contemporary wildlife management and conservation, education, administration, law enforcement, and review articles on the philosophy and history of wildlife management and conservation. This includes: Reports on practices designed to achieve wildlife management or conservation goals. Presentation of new techniques or evaluation of techniques for studying or managing wildlife. Retrospective analyses of wildlife management and conservation programs, including the reasons for success or failure. Analyses or reports of wildlife policies, regulations, education, administration, law enforcement. Review articles on the philosophy and history of wildlife management and conservation. as well as other pertinent topics that are deemed more appropriate for the Wildlife Society Bulletin than for The Journal of Wildlife Management. Book reviews that focus on applied research, policy or wildlife management and conservation.