摘要
由一名新闻记者-机器人与机器学习的工作人员新闻编辑-每日新闻-一项关于人工智能的新研究现在可用。根据NewsRx记者从德国法兰克福发回的新闻报道，研究表明，“许多地区海牛的种群规模基本上还不清楚，主要是因为在浑浊和难以接近的水生环境中进行视觉计数的挑战性。被动声学监测已经显示出在野外监测海牛的前景。”我们的新闻编辑引用了哥特大学Frank Furt的研究：“在这项研究中，我们提出了一种创新的方法，利用传统的神经网络(CNN)来检测、分离和分类来自长期录音的多种声音。为了提高海牛呼叫检测和分类的有效性，CNN分两个阶段工作。首先，将长期录音分成0.5秒的较小窗口，并对其是否包含海牛叫声作出二进制判断。接着，这些发声被分为不同的发声类别(4C类别)，为了分离和分析签名叫声(Squeakes),进一步利用聚类技术来区分被记录的个体并估计种群规模,利用来自三个不同动物设施的不同数量海牛的录音对CNN进行了训练和验证,并对三种不同的聚类方法(两种不同分类器的社区检测和HDBSCAN)进行了适应性测试,结果表明该方法具有良好的适应性。CNN能够准确地检测人声，有效地对不同的呼叫类别进行分类，并证明了HDBSCAN作为聚类方法的可行性。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-A new study on artificial intelligence is now available. According to news reporting originating from Frankfurt am Mai n, Germany, by NewsRx correspondents, research stated, "The population sizes of manatees in many regions remain largely unknown, primarily due to the challengin g nature of conducting visual counts in turbid and inaccessible aquatic environm ents. Passive acoustic monitoring has shown promise for monitoring manatees in t he wild." Our news editors obtained a quote from the research from Goethe-University Frank furt: "In this study, we present an innovative approach that leverages a convolu tional neural network (CNN) for the detection, isolation and classification of m anatee vocalizations from long-term audio recordings. To improve the effectivene ss of manatee call detection and classification, the CNN works in two phases. Fi rst, a longterm audio recording is divided into smaller windows of 0.5 seconds and a binary decision is made as to whether or not it contains a manatee call. S ubsequently, these vocalizations are classified into distinct vocal classes (4 c ategories), allowing for the separation and analysis of signature calls (squeaks ). Signature calls are further subjected to clustering techniques to distinguish the recorded individuals and estimate the population size. The CNN was trained and validated using audio recordings from three different zoological facilities with varying numbers of manatees. Three different clustering methods (community detection with two different classifiers and HDBSCAN) were tested for their suit ability. The results demonstrate the ability of the CNN to accurately detect man atee vocalizations and effectively classify the different call categories. In ad dition, our study demonstrates the feasibility of reliable population size estim ation using HDBSCAN as clustering method."
NETL
NSTL