A novel deep learning based approach for seed image classification and retrieval

•We studied the classification performances of ten different CNN architectures.•We studied and developed a new CNN model, SeedNet, for the classification and retrieval of seed images.•We tested and reported both classification and retrieval results of all the architectures on two very different seed...

Full description

Saved in:

Bibliographic Details
Published in:	Computers and electronics in agriculture Vol. 187; p. 106269
Main Authors:	Loddo, Andrea, Loddo, Mauro, Di Ruberto, Cecilia
Format:	Journal Article
Language:	English
Published:	Amsterdam Elsevier B.V 01-08-2021 Elsevier BV
Subjects:	Agriculture Artificial neural networks Biodiversity Botany Classification Datasets Deep learning Image analysis Image classification Machine learning Object recognition Retrieval Seeds Seeds classification Species classification Deep learning Retrieval Image analysis Agriculture Seeds classification Machine learning
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	•We studied the classification performances of ten different CNN architectures.•We studied and developed a new CNN model, SeedNet, for the classification and retrieval of seed images.•We tested and reported both classification and retrieval results of all the architectures on two very different seeds datasets.•We studied the most appropriate CNN features for seed images retrieval task.•We investigate the extent to which deep feature can be more meaningful with respect to the hand-crafted ones fed to traditional machine learning classifiers. Seeds image analysis has become essential to preserve biodiversity. This is why recognition and classification of plant species on the earth’s planet is nowadays a great challenge. The paper focuses on this purpose by studying two plant seeds datasets to classify their families or species through deep learning techniques. SeedNet, a novel CNN has been proposed to face the depicted issue, and several state-of-the-art convolutional neural networks have been exploited for an exhaustive comparison of most adequate for the considered scenario. In detail, promising results in seed classification for both analysed datasets, reaching accuracy values of 95.65% for the first one and 97.47% for the second one, have been obtained. The retrieval problem with the deep learning approach was also addressed, achieving satisfying performances. We consider the obtained results for both the tasks as an excellent starting point to develop a complete seeds recognition, classification and retrieval system to offer impressive support in agriculture and botany fields.
ISSN:	0168-1699 1872-7107
DOI:	10.1016/j.compag.2021.106269