Actas de congresos
A gravitational model for plant classification using adaxial epidermis texture
Fecha
2015-01-01Registro en:
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 9280, p. 89-96.
1611-3349
0302-9743
10.1007/978-3-319-23234-8_9
2-s2.0-84944790129
Autor
Universidade Federal de Uberlândia (UFU)
Universidade Estadual Paulista (Unesp)
Institución
Resumen
The leaves are very informative plant organs. They are extensively used in plant anatomical studies focusing taxonomy. Their both inner and outer structures provide very discriminant features from vegetal species. In this study, we propose using images from adaxial epidermis for plant classification. The adaxial epidermis is a very variable region in a plant leaf cross-section. It differs in color, number of layers and presence/absence of hypodermis. To accomplish this task, we propose combining complexity analysis methods with a gravitational collapsing system to extract texture features from adaxial epidermis samples. Experimental results show that this combination of techniques surpasses traditional and state-of-the-art methods in both grayscale and color images of adaxial epidermis.