Search

This work proposes an analytical method to identify plant species based on microscopy images of the midrib cross-section of leaves. Unlike previous shape-based approaches based on the individual shape of external contours and cells, an architectural analysis is proposed, where the midrib is semi-automatically segmented and partitioned into histologically relevant structures composed of layers of cells and vascular structures. Using a sequence of morphological operations, a set of geometrical measures from the cells in each layer is extracted to produce a vector of features for species categorization. The method applied to a database containing 10 species of plants from the Brazilian flora achieved a success rate of 91.7%, outperforming other classical shape-based approaches published in the literature. Key words: automatic species identification, image analysis, morphological features. Les auteurs de ce travail proposent une methode analytique permettant d'identifier une espece vegetale a partir d'images en microscopie de coupes transversales de la nervure mediane des feuilles. Contrairement aux approches precedentes basees sur la forme qui reposent sur la forme individuelle des contours externes et sur les cellules, une analyse architecturale est proposee, dans laquelle la nervure mediane est segmentee de maniere semi-automatique et divisee en structures histologiquement pertinentes composees de couches cellulaires et de structures vasculaires. En utilisant une sequence d'operations morphologiques, une serie de mesures geometriques des cellules de chaque couche est extraite afin de produire un ensemble de caracteristiques pour categoriser les especes. La methode appliquee a une base de donnees contenant 10 especes de plantes issues de la flore du Bresil a obtenu un taux de succes de 91.7%, surpassant les autres approches classiques basees sur la forme publiees dans la litterature. [Traduit par la Redaction] Mots-cles : identification automatisee d'une espece, analyse d'image, caracteristiques morphologiques., Introduction The categorization of plants into species is an important taxonomical task; it provides meaningful information about particular characteristics of a species, which can be used in the context of [...]

Background: Primordial odontogenic tumour is a benign mixed neoplasm of recent description, which has histological similarities with other odontogenic tumours such as the ameloblastic fibroma. In this article, we investigate the architecture of the sub‐epithelial layer of mesenchymal cells expressing the marker CD34 in primordial odontogenic tumour. Objective: Analyse the spatial patterns of CD34 expression in primordial odontogenic tumour and compare them with those in ameloblastic fibroma and the normal tooth germ by means of objective imaging approaches, to better characterise these lesions. Methods: Two cases of primordial odontogenic tumour, four cases of ameloblastic fibroma and two cases of tooth germ in cap and bell stages were used for morphological, structural and immunohistochemical analyses. Results: CD34 expression was found in vascular endothelium of primordial odontogenic tumour, ameloblastic fibroma and tooth germ. In addition, a characteristic sub‐epithelial expression was observed only in primordial odontogenic tumour, corresponding to 84%–86% of the sample boundaries. Moreover, the zone expressing CD34 corresponded with a higher cellularity, which was absent in ameloblastic fibroma and tooth germ. Conclusion: Image analysis of the primordial odontogenic tumour architecture revealed characteristics absent in other odontogenic tumours and tooth germs. This study provides additional information to support the idea that this neoplasm is a distinct entity from early stage AF or developing odontoma. [ABSTRACT FROM AUTHOR]