Flower Dataset Research Articles

YNU Flower: A market-oriented benchmark dataset for flower classification

The great cultural values of flowers in our life have nourished a highly developed flower industry. However, our current flower trade is still labour-intensive mainly due to a lack of effective automated classification systems. While several image datasets have been established and are widely used for developing visual flower classification models, they are of little practical value to the industry as, in these datasets, few flowers of market value are included, and intra-class and temporal variabilities are also not taken into consideration. Here, we present a large-scale flower image dataset named YNU Flower for flower classification. A total of 91 commercial flower varieties belonging to eleven favoured botanical groups were purchased from the Dounan flower market, the largest flower trading centre in Asia, and 17,806 images were captured. Notably, the YNU Flower dataset includes many commercial varieties with subtle appearance distinctions but significantly different price tags. Using this newly established dataset, we trained several representative models for classification and object detection tasks. By making YNU Flower publicly available to the community, we anticipate that this benchmark dataset will expedite translational research on flower classification and ultimately benefit the flower industry.

Deep learning-based restoration of nonlinear motion blurred images for plant classification using multi-spectral images

There have been various plant image-based studies for segmentation, deblurring, super-resolution reconstruction, and classification. However, nonlinear motion blur in thermal images was not considered in the existing studies on plant classification. Nonlinear motion blur occurs in images due to camera or plant movements, and it causes the degradation of plant classification accuracy. Moreover, nonlinear motion blur in images gets worse when both camera and plant movements occur simultaneously. In this case, it becomes difficult to recognize plants, and the performance of plant image classification becomes very low. Therefore, to reduce the nonlinear motion blur, a thermal and visible light plant images-based deblurring network (TVPD-Net) is proposed in this study. In addition, a thermal and visible light plant images-based classification network (TVPC-Net) is also proposed to improve the plant classification performance on deblurred images. Experimental results revealed that the proposed TVPD-Net achieved 21.21 and 22.53 of the peak signal-to-noise ratio (PSNR), and 0.726 and 0.737 of the structural similarity index measure (SSIM) on both visible light and thermal plant image datasets which were self-collected, respectively. Moreover, the proposed TVPC-Net with deblurred images by TVPD-Net achieved 92.52 % (top-1 accuracy) and 87.73 % (harmonic mean of precision and recall (F1-score)). In addition, the experimental results on an open dataset named Hyperspectral Flower Dataset (HFD100) revealed that the proposed plant classification method achieved 90.94 % of top-1 accuracy and 86.21 % of F1-score. The accuracies of the proposed methods are greater than those of the state-of-the-art methods.

Weakly supervised classification through manifold learning and rank-based contextual measures

Over the last decade, significant advances have been achieved by machine learning approaches, notably in supervised learning scenarios. Supported by the advent of deep learning and comprehensive training sets, the accuracy achieved on classification tasks has improved significantly. Simultaneously, we have experienced massive growth in multimedia data and applications, which have become ubiquitous in several domains. However, with the increase in multimedia data collections, significant bottlenecks associated with the lack of labeled data emerged. To surpass this critical issue, developing methods capable of exploiting the unlabeled data and operating under weak supervision has become imperative. This work proposes a rank-based model capable of using contextual information encoded in the unlabeled data to perform weakly supervised classification. We evaluated the proposed weakly supervised approach on multimedia classification tasks with and without manifold learning algorithms, considering several combinations of rank correlation measures and classifiers. An experimental evaluation was conducted on 6 public image datasets considering different features, including convolutional neural networks and visual transformers. Positive gains were achieved compared to supervised and semi-supervised baselines for the same amount of labeled data. For instance, the proposed approach with manifold learning enhanced the accuracy of the Optimum-Path Forest (OPF) classifier from 71.77% to 83.24% when applied to the Flowers dataset and ResNet features. Among the conclusions, this work reveals that rank-based correlation measures and manifold learning can be used for a more effective labeled set expansion.

Regularization Effects in Deep Learning Architecture

This research examines the impact of three widely utilized regularization approaches -- data augmentation, weight decay, and dropout --on mitigating overfitting, as well as various amalgamations of these methods. Employing a Convolutional Neural Network (CNN), the study assesses the performance of these strategies using two distinct datasets: a flower dataset and the CIFAR-10 dataset. The findings reveal that dropout outperforms weight decay and augmentation on both datasets. Additionally, a hybrid of dropout and augmentation surpasses other method combinations in effectiveness. Significantly, integrating weight decay with dropout and augmentation yields the best performance among all tested method blends. Analyses were conducted in relation to dataset size and convergence time (measured in epochs). Dropout consistently showed superior performance across all dataset sizes, while the combination of dropout and augmentation was the most effective across all sizes, and the triad of weight decay, dropout, and augmentation excelled over other combinations. The epoch-based analysis indicated that the effectiveness of certain techniques scaled with dataset size, with varying results.

Further exploration on deep learning in flower recognition

Flower recognition is an important research direction in the field of computer vision, and automatic classification of flower images through deep learning methods is of great significance for ecological environment monitoring and plant research. With this background, this study aims to further optimize the existing flower recognition system and improve its classification accuracy by adjusting key parameters. Based on the existing deep learning model and several rounds of training, this paper explores the tuning strategies for parameters such as different learning rates and weight decay to achieve higher recognition accuracy. In the experiments, this paper uses the classical flower dataset and enhances the diversity of the data through image preprocessing and data enhancement. Through a hundred rounds of training, the model in this paper achieves about 80% classification accuracy on the test set, which is significantly improved compared with the initial model. Further analysis of the results shows that by reasonably adjusting the learning rate and weight decay parameters, the model achieves certain improvements in different flower classes, demonstrating the impact of parameter tuning on the overall performance of the model.

Potted Phalaenopsis Grading: Precise Bloom and Bud Counting with the PA-YOLO Algorithm and Multiviewpoint Imaging

The accurate detection and counting of flowers ensure the grading quality of the ornamental plants. In automated potted flower grading scenarios, low detection precision, occlusions and overlaps impact counting accuracy. This study proposed a counting method combining a deep learning algorithm with multiple viewpoints. Firstly, a flower detection model, PA-YOLO, was developed based on YOLOv5 by designing a two-scale detection branch, optimizing the number of bottlenecks and integrating a dynamic head framework. Next, PA-YOLO was used to detect grouped 360-viewpoint images of each potted plant to determine the optimal number of viewpoints for counting. The detection results indicated that PA-YOLO achieved a mean average precision (mAP) of 95.4% and an average precision (AP) of 91.9% for occluded blooms on our Phalaenopsis flower dataset. For the optimal number of viewpoints, the average counting accuracy of buds and blooms was highest at three viewpoints, with scores of 96.25% and 93.33%, respectively. The final counting accuracy reached 95.56% in flower counting tests conducted from three viewpoints. The overall results suggest that the proposed method can effectively detect and count flowers in complex occlusion and overlap environments, providing guidance for designing and implementing the vision component in an automated potted flower grading system.

A Framework for Single-Panicle Litchi Flower Counting by Regression with Multitask Learning

The number of flowers is essential for evaluating the growth status of litchi trees and enables researchers to estimate flowering rates and conduct various phenotypic studies, particularly focusing on the information of individual panicles. However, manual counting remains the primary method for quantifying flowers, and there has been insufficient emphasis on the advancement of reliable deep learning methods for estimation and their integration into research. Furthermore, the current density map-based methods are susceptible to background interference. To tackle the challenges of accurately quantifying small and dense male litchi flowers, a framework counting the flowers in panicles is proposed. Firstly, an existing effective algorithm YOLACT++ is utilized to segment individual panicles from images. Secondly, a novel algorithm FlowerNet based on density map regression is proposed to accurately count flowers in each panicle. By employing a multitask learning approach, FlowerNet effectively captures both foreground and background information, thereby overcoming interference from non-target areas during pixel-level regression tasks. It achieves a mean absolute error of 47.71 and a root mean squared error of 61.78 on the flower dataset constructed. Additionally, a regression equation is established using a dataset of inflorescences to examine the application of the algorithm for flower counting. It captures the relationship between the predicted number of flowers by FlowerNet and the manually counted number, resulting in a determination coefficient ( R 2 ) of 0.81. The proposed algorithm shows promise for automated estimation of litchi flowering quantity and can serve as a valuable reference for litchi orchard management during flowering period.

OPTIMIZING LANTANA CLASSIFICATION: HIGH-ACCURACY MODEL UTILIZING FEATURE EXTRACTION

As an invasive and poisonous plant, Lantana has become a pest in the agricultural world. Still, on the other hand, it becomes an ornamental plant with different positive potentials. Lantana flower datasets are not yet widely available for open image classification research, given that the research needs are still broad in remote sensing. This study aims to provide a model with classifier accuracy that outperforms similar studies and Lantana datasets for classification needs using several algorithms that can be run on small source computers. This study used five types of lantana colors, red, white, yellow, purple, and orange, as the primary dataset, which had 411 instances. VGG16 assisted feature extraction in preparing datasets for the data training using three classifiers: decision tree, AdaBoost, and k-NN. 2-fold cross-validation, 5-fold cross-validation, and a self-organizing map are used to help validate each process. The experiment to measure the classifier's performance resulted in a good figure of 99.8% accuracy for 2-fold cross-validation, 100% for 5-fold cross-validation, and a primary dataset of lantana interest that can be accessed freely on the IEEE Data port. This study outperformed other related studies in terms of classifier accuracy.

Implementation of Flower Recognition using Convolutional Neural Networks

The recognition of flowers holds significant importance within the realms of ecological research, horticulture, and diverse technological applications. This study presents "Blossom Insight," an innovative methodology for flower identification that employs Convolutional Neural Networks within the Keras framework. This study aims to examine the necessity of precise and effective flower categorization, considering the extensive range of floral species. The methodology encompasses a rigorous procedure of data preprocessing, utilizing sophisticated techniques to augment the model's capacity to distinguish intricate characteristics of flowers. The crux of the study revolves around the amalgamation of a Convolutional Neural Network, a robust deep learning methodology, with Keras, a user-accessible open-source framework for machine learning. The integration of these components enables the development of a resilient flower recognition model that possesses the ability to acquire complex patterns and characteristics from input images. The training of the model encompasses exposure to a wide range of flower datasets, which enhances its ability to generalize across different species and environmental conditions effectively. The findings illustrate the effectiveness of "Blossom Insight" in attaining a notable level of precision in tasks related to the identification of flowers. The implementation not only contributes to the advancement of the field of computer vision but also offers a valuable resource for researchers, horticulturists, and enthusiasts seeking a comprehensive understanding and accurate identification of floral species. The development of "Blossom Insight" signifies a notable advancement in utilizing deep learning techniques to augment our understanding and admiration of the wide variety present in the realm of flowers.

Variational quantum state discriminator for supervised machine learning

Abstract Quantum state discrimination (QSD) is a fundamental task in quantum information processing with numerous applications. We present a variational quantum algorithm that performs the minimum-error QSD, called the variational quantum state discriminator (VQSD). The VQSD uses a parameterized quantum circuit that is trained by minimizing a cost function derived from the QSD, and finds the optimal positive-operator valued measure (POVM) for distinguishing target quantum states. The VQSD is capable of discriminating even unknown states, eliminating the need for expensive quantum state tomography. Our numerical simulations and comparisons with semidefinite programming demonstrate the effectiveness of the VQSD in finding optimal POVMs for minimum-error QSD of both pure and mixed states. In addition, the VQSD can be utilized as a supervised machine learning algorithm for multi-class classification. The area under the receiver operating characteristic curve obtained in numerical simulations with the Iris flower dataset ranges from 0.97 to 1 with an average of 0.985, demonstrating excellent performance of the VQSD classifier.

Anchor-free deep convolutional neural network for tracking and counting cotton seedlings and flowers

Accurate counting of plants and their organs in natural environments is essential for breeders and growers. For breeders, counting plants during the seedling stage aids in selecting genotypes with superior emergence rates, while for growers, it informs decisions about potential replanting. Meanwhile, counting specific plant organs, such as flowers, forecasts yields for different genotypes, offering insights into production levels. The overall goal of this study was to investigate a deep convolutional neural network-based tracking method, CenterTrack, for cotton seedling and flower counting from video frames. The network is extended from a customized CenterNet, which is an anchor-free object detector. CenterTrack predicts the detections of the current frame and displacements of detections between the previous frame and the current frame, which are used to associate the same object in consecutive frames. The modified CenterNet detector achieved high accuracy on both seedling and flower datasets with an overall AP50 of 0.962. The video tracking hyperparameters were optimized for each dataset using orthogonal tests. Experimental results showed that seedling and flower counts with optimized hyperparameters highly correlated with those of manual counts (R2 = 0.98 andR2 = 0.95) and the mean relative errors of 75 cotton seedling testing videos and 50 flower testing videos were 5.5 % and 10.8 %, respectively. An average counting speed of 20.4 frames per second was achieved with an input resolution of 1920 × 1080 pixels for both seedling and flower videos. The anchor-free deep convolution neural network-based tracking method provides automatic tracking and counting in video frames, which will significantly benefit plant breeding and crop management.

SMART AND INTELLIGENT PRODUCTION STRATEGY FOR THE FLOWER MARKET USING DATA MINING KNOWLEDGE-BASED DECISION

The agriculture industry has been an enormous economic pillar in the production and consumption market value chain. The agriculture industry resets flower production factors with the agricultural technology revolution. The fastest technology provides innovative and intelligent decision-making strategies in seasonal cut flowers to increase production. This study briefs out existing farming practices, chain activity and farming technology’s significant impacts on the agriculture field and garden industry. Authors try to investigate cut flower production status and analyze production values to design innovative and intelligent strategies, especially for seasonal flower production. The study employs a flower dataset; hence, it applies floral parameter inputs and data mining association rules to create an output of the flower production category, which fits appropriately to evaluate flower market production value in a particular season. The article's result reveals that the proposed flower production strategy provides efficient and intelligent guidelines to increase flower production according to market demand. This study suggests an intelligent and friendly production strategy for gardeners that indicates the flower market gets continuous and quality production to meet consumers’ immediate market demand.

A New Procedure for Unsupervised Clustering Based on Combination of Artificial Neural Networks

Classification methods have become one of the main tools for extracting essential information from multivariate data. New classification algorithms are continuously being proposed and created. This paper presents a classification procedure based on a combination of Kohonen and probabilistic neural networks. Its applicability and efficiency are estimated using model data sets (iris flowers data set, wine data set, data with a two-hierarchical structure), then compared with the traditional clustering algorithms (hierarchical clustering, k-means clustering, fuzzy k-means clustering). The algorithm was designed as M-script in Matlab 7.11b software. It was shown that the proposed classification procedure has a great advantage over traditional clustering methods.

The effectiveness of T5, GPT-2, and BERT on text-to-image generation task

Text-to-image (T2I) generation, which involves synthesizing an image from a textual description, has emerged as a popular research topic in computer vision. Meanwhile, transformer-based models, such as BERT, GPT-2, and T5, have demonstrated promising results in various natural language processing tasks, including text generation and translation. However, their application to T2I generation remains largely unexplored. Therefore, a comparative study investigating the performance of BERT, GPT-2, and T5 in T2I generation is of significant importance. Such a study would shed light on the strengths and weaknesses of each model, facilitating the identification of the most suitable approach for this task. In this paper, we propose three architectures to conduct a comparative study of T5, GPT-2, and BERT in T2I generation tasks. We fine-tune these models to generate text vectors and then transform the textual information into images using the affine transformation into the DF-GAN generator. Subsequently, we evaluate the quality, diversity, and the models’ ability to recognize the words. Our experiments on the challenging CUB and Oxford-102 flower datasets demonstrate that T5 exhibits promising potential for T2I generation. It has the capability to generate visually appealing and semantically coherent images from textual descriptions.

Radish Growth Stage Recognition Based on GAN and Deep Transfer Learning

Image recognition of plant growth states provides technical support for crop monitoring; this reduces labor costs and promotes efficient planting. However, difficulties in data collection, the required high levels of algorithm efficiency, and the lack of computing power resources create challenges to the development of intelligent agriculture. As a result, a deep transfer learning algorithm is proposed in this paper. The main motivation for this study was the lack of a dataset of plant growth stages. The key idea was to collect radish growth stage images in an experimental field using standardized equipment and to generate more images using DCGAN. By improving the deep transfer learning model, radish growth stages can be identified much more accurately. In this study, five different deep migration models were selected, namely, Inception-v3, MobileNet, Xception, VGG-16, and VGG-19. Our experiment demonstrated that Inception-v3 was the most suitable model for the recognition of plant growth states. Based on Inception-v3, we propose three improved models. The test accuracies for the radish and Oxford Flower datasets were 99.5% and 99.3%, respectively. Additionally, the accuracy of the pest and disease dataset also achieved excellent performance, with an accuracy of 94.7%, 2.4% higher than previously. These results demonstrate the wide applicability of our model and the rationality of constructing a radish growth stage dataset.

Zero-Shot Image Classification with Rectified Embedding Vectors Using a Caption Generator

Although image recognition technologies are developing rapidly with deep learning, conventional recognition models trained by supervised learning with class labels do not work well when test inputs from untrained classes are given. For example, a recognizer trained to classify Asian bird species cannot recognize the species of kiwi, because the class label “kiwi” and its image samples have not been seen during training. To overcome this limitation, zero-shot classification has been studied recently, and the joint-embedding-based approach has been suggested as one of the promised solutions. In this approach, image features and text descriptions belonging to the same class are trained to be closely located in a common joint-embedding space. Once we obtain the embedding function that can represent the semantic relationship of image–text pairs in training data, test images and text descriptions (prototypes) of unseen classes can also be mapped to the joint-embedding space for classification. The main challenge with this approach is mapping inputs of two different modalities into a common space, and previous works suffer from the inconsistency between the distribution of two feature sets on joint-embedding space extracted from the heterogeneous inputs. To treat this problem, we propose a novel method of employing additional textual information to rectify the visual representation of input images. Since the conceptual information of test classes is generally given as texts, we expect that the additional descriptions from a caption generator can adjust the visual feature for better matching with the representation of the test classes. We also propose to use the generated textual descriptions to augment training samples for learning joint-embedding space. In the experiments on two benchmark datasets, the proposed method shows significant performance improvements of 1.4% on the CUB dataset and 5.5% on the flower dataset, in comparison to existing models.

Hybridizing grid partitioning, rough set theory, and feature selection for fuzzy rule generation in dataset classification

This research investigates the hybridization of Grid Partitioning, Rough Set Theory, and Feature Selection for Fuzzy Rule Generation in Dataset Classification. The objective is to improve classification accuracy and interpretability by integrating multiple techniques. Grid partitioning is employed to divide the dataset into regions, allowing localized analysis. Rough set theory is utilized for attribute reduction and feature selection, identifying informative features within each region. Fuzzy rule generation is applied to generate interpretable classification rules using linguistic terms and membership functions. The hybrid model is optimized using metaheuristic algorithms to maximize classification performance. The research demonstrates the potential of the hybrid approach through experiments on the Iris flower dataset. The findings reveal improved classification accuracy, enhanced interpretability, and effective handling of complex datasets. The research contributes to the field by integrating these techniques into a cohesive framework and highlights the importance of parameter settings, computational complexity, and real-world applications. Future work should address these limitations and validate the approach on diverse datasets. The hybridization of Grid Partitioning, Rough Set Theory, and Feature Selection for Fuzzy Rule Generation holds promise for advancing classification models in various domains

Quantifying Cell-Type-Specific Differences of Single-Cell Datasets Using Uniform Manifold Approximation and Projection for Dimension Reduction and Shapley Additive exPlanations.

With rapid advances in single-cell profiling technologies, larger-scale investigations that require comparisons of multiple single-cell datasets can lead to novel findings. Specifically, quantifying cell-type-specific responses to different conditions across single-cell datasets could be useful in understanding how the difference in conditions is induced at a cellular level. In this study, we present a computational pipeline that quantifies cell-type-specific differences and identifies genes responsible for the differences. We quantify differences observed in a low-dimensional uniform manifold approximation and projection for dimension reduction space as a proxy for the difference present in the high-dimensional space and use SHapley Additive exPlanations to quantify genes driving the differences. In this study, we applied our algorithm to the Iris flower dataset, single-cell RNA sequencing dataset, and mass cytometry dataset and demonstrate that it can robustly quantify cell-type-specific differences and it can also identify genes that are responsible for the differences.

Nano Aerial Vehicles for Tree Pollination

Currently, Unmanned Aerial Vehicles (UAVs) are considered in the development of various applications in agriculture, which has led to the expansion of the agricultural UAV market. However, Nano Aerial Vehicles (NAVs) are still underutilised in agriculture. NAVs are characterised by a maximum wing length of 15 centimetres and a weight of fewer than 50 g. Due to their physical characteristics, NAVs have the advantage of being able to approach and perform tasks with more precision than conventional UAVs, making them suitable for precision agriculture. This work aims to contribute to an open-source solution known as Nano Aerial Bee (NAB) to enable further research and development on the use of NAVs in an agricultural context. The purpose of NAB is to mimic and assist bees in the context of pollination. We designed this open-source solution by taking into account the existing state-of-the-art solution and the requirements of pollination activities. This paper presents the relevant background and work carried out in this area by analysing papers on the topic of NAVs. The development of this prototype is rather complex given the interactions between the different hardware components and the need to achieve autonomous flight capable of pollination. We adequately describe and discuss these challenges in this work. Besides the open-source NAB solution, we train three different versions of YOLO (YOLOv5, YOLOv7, and YOLOR) on an original dataset (Flower Detection Dataset) containing 206 images of a group of eight flowers and a public dataset (TensorFlow Flower Dataset), which must be annotated (TensorFlow Flower Detection Dataset). The results of the models trained on the Flower Detection Dataset are shown to be satisfactory, with YOLOv7 and YOLOR achieving the best performance, with 98% precision, 99% recall, and 98% F1 score. The performance of these models is evaluated using the TensorFlow Flower Detection Dataset to test their robustness. The three YOLO models are also trained on the TensorFlow Flower Detection Dataset to better understand the results. In this case, YOLOR is shown to obtain the most promising results, with 84% precision, 80% recall, and 82% F1 score. The results obtained using the Flower Detection Dataset are used for NAB guidance for the detection of the relative position in an image, which defines the NAB execute command.

Modulating STDP With Back-Propagated Error Signals to Train SNNs for Audio Classification

Audio classification has many practical applications such as noise pollution detection, wildlife monitoring, speech recognition, and more. For many of these applications, deploying classifiers on low powered devices for persistent deployment is desirable. Artificial neural networks (ANN) have achieved state-of-the-art performance on audio classification tasks; however, it is not always feasible to deploy modern ANNs to embedded devices due to their high power consumption. Biologically inspired spiking neural networks (SNN) have been shown to significantly reduce power consumption during inference when compared with equivalent ANNs, and they have also been theoretically proven to be more computationally powerful than stateless ANNs. This work proposes an audio classification system using SNNs, and a learning algorithm is developed for classification with multilayer SNNs which combines biologically plausible spike-timing-dependent plasticity (STDP) with spatial error backpropagation. By allowing the STDP process to account for both temporal dependencies and the non-differentiable activation function derivative, the proposed learning rule successfully trains multilayer SNNs for the considered classification tasks. Through the STDP process, the proposed learning rule is also capable of online learning; explicit storage of values from previous timesteps is not required, in contrast to the widely adopted backpropagation through time (BPTT) algorithm. The proposed approach approaches the performance of SNNs trained via BPTT on the classification tasks. SNNs trained with the proposed learning rule are evaluated on the Iris Flower dataset, the Real-World Computing Partnership sounds dataset, the Free Spoken Digits Dataset, and the UrbanSound8k dataset.