Sample Entropy Signatures: A new way to interpret SampEn values[Formula presented]

• Autor(a) principal: Rozendo, Guilherme Botazzo [UNESP]
• Data de Publicação: 2022
• Outros Autores: do Nascimento, Marcelo Zanchetta, Roberto, Guilherme Freire, de Faria, Paulo Rogério, Silva, Adriano Barbosa, Tosta, Thaína Aparecida Azevedo, Neves, Leandro Alves [UNESP]
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• DOI: 10.1016/j.simpa.2022.100329
• Texto Completo: http://dx.doi.org/10.1016/j.simpa.2022.100329; http://hdl.handle.net/11449/241203
• topic: Feature extraction; Pattern recognition; SampEn signatures; Sample entropy; Signature metrics; Texture signatures

We present the Sample Entropy Signatures method, which provides a new way to represent images via Sample entropy (SampEn). The proposed method defines texture signatures from multiple SampEn values obtained from combinations of the parameters m and r. Here, m signatures were determined (m ranging from 1 to 4), with r points each (r ranging from 0.06 to 0.4), and the behavior of each signature was determined by the metrics: area under the curve, skewness, maximum entropy value, and area ratio. This approach aimed to improve the quantitative capacity of SampEn and provide a new interpretation of its values.