Sift and Deep Convolutional Features for Closeness-Based Leaf Image Recognition

Sift and Deep Convolutional Features for Closeness-Based Leaf Image Recognition

Sucithra B. (Anna University, Chennai, India) and Angelin Gladston (College of Engineering, Anna University, Chennai, India)
Copyright: © 2020 |Pages: 14
DOI: 10.4018/IJCVIP.2020040102


Plant leaf recognition has been carried out widely using low level features. Scale invariant feature transform techniques have been used to extract the low level features. Leaves that match based on low level features but do not do so in the semantic perspective cannot not be recognized. To address that, global features have been extracted and used using convolutional neural networks. Even then there are issues like leaf images in various illuminations, rotations, taken in different angles, and so on. To address such issues, the closeness among low level features and global features are computed using multiple distance measures and a leaf recognition framework has been proposed. The matched patches are evaluated both quantitatively and qualitatively. Experimental results obtained are promising for the proposed closeness-based leaf recognition framework.
Article Preview


There are a wide variety of plant species, many of them are useful to humans as food or as medicine (Chithra & Janes, 2018), some are close to extinction and few others are harmful to man. Apart from this, plants are vital in their role they are not only essential for human beings; they are also the base of all food chains, being the producers of food. To protect such plants and sustain biodiversity, we need to acquire in-depth knowledge on how to use and protect these plant species (Guoqing et al., 2019), in spite of the existing challenges (Erick & Jose, 2018) in studying and classifying these plants correctly. First and foremost is identifying the unknown plants which mainly depends on the expertise gained by an expert botanist (Guoqing et al., 2019). Traditionally, the successful method to identify plants easily, is by using the manual-based method based on their morphological characteristics. The success behind using this method for classifying the plant species is mainly rooted on the acquired knowledge and human skills. Plant components such as flowers, fruits, stem, seeds, root and leaves are used in plant identification and classification (Jibi et al., 2018). Predominantly leaf has been used for plant recognition, since leaves stay on the plants for more months. The leaves look completely different from one another exhibiting various characteristics (Neha et al., 2018) such as color, size, kind like maple, and oak, number of points, as well as arrangement of veins. Different plant species contains different leaf characteristics.

Botanists use their knowledge on leaves to identify the plant species and classify them as dangerous species, species having medicinal purposes as well as plant species that are edible. However, this process of manual-based leaf recognition had been often laborious and also consumes more time for leaf recognition. Hence, researchers have conducted studies to support the automatic classification of plants based on their physical characteristics such as the color, shape (Feng & Bin, 2018) and size (Cao et al., 2017). Many leaf recognition systems were developed following the sequence of processing steps, namely preparing the leaf dataset, pre-processing to extract their specific features, classification of the leaves, populating the database, training for recognition and finally evaluating the results. Such automatic leaf classification of plants uses various machine learning algorithms (Affix et al., 2018; Zhou et al., 2016). Leaf features can be categorized into low level and high level features. Low level leaf features are the boundary, color, illumination, as well as scaling features of the leaf and high level leaf features are the shape descriptors, texture features, length, leaf tips, color representations, venation structure, and morphology namely, the form and structure of the leaf.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 11: 4 Issues (2021): Forthcoming, Available for Pre-Order
Volume 10: 4 Issues (2020): 2 Released, 2 Forthcoming
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 2 Issues (2016)
Volume 5: 2 Issues (2015)
Volume 4: 2 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing