Metric learning for image-based flower cultivars identification

Abstract

Background The study of plant phenotype by deep learning has received increased interest in recent years, which impressive progress has been made in the fields of plant breeding. Deep learning extremely relies on a large amount of training data to extract and recognize target features in the field of plant phenotype classification and recognition tasks. However, for some flower cultivars identification tasks with a huge number of cultivars, it is difficult for traditional deep learning methods to achieve better recognition results with limited sample data. Thus, a method based on metric learning for flower cultivars identification is proposed to solve this problem. Results We added center loss to the classification network to make inter-class samples disperse and intra-class samples compact, the script of ResNet18, ResNet50, and DenseNet121 were used for feature extraction. To evaluate the effectiveness of the proposed method, a public dataset Oxford 102 Flowers dataset and two novel datasets constructed by us are chosen. For the method of joint supervision of center loss and L2-softmax loss, the test accuracy rate is 91.88%, 97.34%, and 99.82% across three datasets, respectively. Feature distribution observed by T-distributed stochastic neighbor embedding (T-SNE) verifies the effectiveness of the method presented above. Conclusions An efficient metric learning method has been described for flower cultivars identification task, which not only provides high recognition rates but also makes the feature extracted from the recognition network interpretable. This study demonstrated that the proposed method provides new ideas for the application of a small amount of data in the field of identification, and has important reference significance for the flower cultivars identification research.

Breakdancing Movement Based on Image Recognition Promotes Preschool Children’s Executive Function and Intervention Plan

“…Zhang et al have described an effective metric learning method for the task of identifying flower varieties. It not only provides a high recognition rate but also makes the features extracted from the recognition network interpretable [5]. Aprahamian challenges this narrative by introducing the voices of the founding practitioners that break is often overlooked.…”

Application of Image Super-Resolution Recognition and Artificial Intelligence System in Repairing Students’ Psychological Education Problems

“…It can be seen that the grayscale information of the SAR image filtered by this algorithm is the closest to the original image, and there is no obvious noise peak in the grayscale value. It also can be seen that the filtering effect of the improved denoising algorithm in this paper is better than other methods [15]. If there is relative motion between the camera and the scene, the first step of super-resolution reconstruction is to register multiple frames, that is, to calculate the pixel displacement of the reference frame image relative to other images.…”

Comparing Deep Learning Approaches for Understanding Genotype × Phenotype Interactions in Biomass Sorghum

“…Many approaches to solve this problem have been proposed in recent years, both varying specific loss functions to define the embedding (Hadsell et al, 2006 ; Sohn, 2016 ; Ge, 2018 ; Kim et al, 2018 ; Xuan et al, 2018 ), and proposing interesting datasets along with loss functions (Schroff et al, 2015 ; Song et al, 2016 ). In this work we use a variation called the Proxy Loss approach described in (Movshovitz-Attias et al ( 2017 ) and Boudiaf et al ( 2020 ), which was recently used for plant-recognition based on flower images (Zhang et al, 2021 ). This trains an embedding network so that images taken from the same field plot are mapped closer together than images taken from different field plots; this source of weak labeling would apply to any situation where field plots consist of unique cultivars.…”