Multi-algorithm Respiratory Crackle Detection
Multi-algorithm Respiratory Crackle Detection
复制标题
多算法呼吸爆裂检测
DOI：
--
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
A. Marques
中科院分区：
文献类型：
--
作者：
J. Quintas;Guilherme Campos;A. Marques
基于文献中提出的技术，实现了四种裂纹检测算法。这些算法在一组肺音上进行了测试，并根据灵敏度（SE）、准确性（PPV）及其谐波平均值（F指数）对其性能进行了评估。计算这些指标的参考注释数据是通过三名卫生专业人员对同一组肺音的独立注释的多数一致获得的。四种算法中的大多数的一致性比最佳单个算法提供了7%以上的性能改进。
Four crackle detection algorithms were implemented based on selected techniques proposed in the literature. The algorithms were tested on a set of lung sounds and their performance was assessed in terms of sensitivity (SE), accuracy (PPV) and their harmonic mean (F index). The reference annotation data for calculating these indices were obtained through agreement by majority between independent annotations made by three health professionals on the same set of lung sounds. Agreement by majority of the four algorithms afforded more than 7% performance improvement over the best individual algorithm.