H-Metric: Characterizing Image Datasets via Homogenization Based on KNN-Queries

Data Science Journal

Sergio F. da Silva²  Agma J. M. Traina²  Welington M. da Silva¹  Jose F. Rodrigues Jr.² 

[1] Universidade Federal de Sao Carlos - Campus Sorocaba - Rodovia Joao Leme dos Santos;Inst. de Ciencias Matematicas e de Computacao - Universidade de Sao Paulo

关键词: Content-based image retrieval;    Metric spaces;    Precision-Recall;   

DOI  :  10.2481/dsj.10-007

学科分类：计算机科学（综合）

来源: Ubiquity Press Ltd.

【 摘 要 】

References(8)Precision-Recall is one of the main metrics for evaluating content-based image retrieval techniques. However, it does not provide an ample perception of the properties of an image dataset immersed in a metric space. In this work, we describe an alternative metric named H-Metric, which is determined along a sequence of controlled modifications in the image dataset. The process is named homogenization and works by altering the homogeneity characteristics of the classes of the images. The result is a process that measures how hard it is to deal with a set of images in respect to content-based retrieval, offering support in the task of analyzing configurations of distance functions and of features extractors.

【 授权许可】

Unknown