Deep neural networks for explainable feature extraction in orchid identification

Abstract

Automated image-based plant identification systems are black-boxes, failing to provide an explanation of a classification. Such explanations are seen as being essential by taxonomists and are part of the traditional procedure of plant identification. In this paper, we propose a different method by extracting explicit features from flower images that can be employed to generate explanations. We take the benefit of feature extraction derived from the taxonomic characteristics of plants, with the orchids as an example domain. Feature classifiers were developed using deep neural networks. Two different methods were studied: (1) a separate deep neural network was trained for every individual feature, and (2) a single, multi-label, deep neural network was trained, combining all features. The feature classifiers were tested in predicting 63 orchid species using naive Bayes (NB) and tree-augmented Bayesian networks (TAN). The results show that the accuracy of the feature classifiers is in the range 83-93%. By combining these features using NB and TAN the species can be predicted with an accuracy of 88.9%, which is better than a standard pre-trained deep neural-network architecture, but inferior to a deep learning architecture after fine-tuning of multiple layers. The proposed novel feature extraction method still performs well for identification and is explainable, as opposed to black-box solutions that only aim for the best performance. Graphical abstract

Use of remote sensing and image processing for identification of wild orchids

“…Other researchers have applied machine learning (Sabri et al, 2019;Andono et al, 2021) or neural network algorithms (Arwatchananukul et al, 2020;Apriyanti et al, 2021;Sarachai et al, 2022;Apriyanti et al, 2023;Ou et al, 2023) to extract flower features (e.g., colour, shape) from the library images of orchids. These approaches were also able to obtain high accuracies (82%-99%).…”

“…Further, several other studies utilised deep neural networks to classify cultivated orchid species (Arwatchananukul et al, 2020;Sarachai et al, 2022;Ou et al, 2023). A recent study used taxonomic features, which are additional botanic features apart from color, shape, and texture, to classify 63 cultivated orchid species using naïve Bayes and treeaugmented Bayesian networks (TAN) (Apriyanti et al, 2023). The accuracy of this classification algorithm to classify the orchid species was recorded as 89%.…”

“…While only a limited number of studies on orchid species have been reported, the high accuracy of classification reported shows considerable promise for the development and application of OBIA methods. However, it must be noted that the reported orchid species classification approaches (Sabri et al, 2019;Arwatchananukul et al, 2020;Andono et al, 2021;Sarachai et al, 2022;Apriyanti et al, 2023;Ou et al, 2023) have all been done on cultivated orchids and not applied to species found in the wild.…”