Article
Plant Sciences
Yanping Zhang, Jing Peng, Xiaohui Yuan, Lisi Zhang, Dongzi Zhu, Po Hong, Jiawei Wang, Qingzhong Liu, Weizhen Liu
Summary: A novel cultivar identification system called MFCIS was proposed, which achieved high accuracy by combining persistent homology and CNN to extract multiple leaf morphological features. The system was validated on sweet cherry and soybean datasets, demonstrating an effective method for plant breeders to identify plant cultivars efficiently.
HORTICULTURE RESEARCH
(2021)
Article
Computer Science, Information Systems
Yi-Lin Tu, Wei-Yang Lin, Yao-Cheng Lin
Summary: The development of automatic plant phenotyping systems has gained great attention recently due to its ability to improve measurement efficiency and reduce labor costs. In this study, we focus on developing an automatic method for leaf counting by treating it as an object detection problem. By utilizing the YOLOv3 network model, we achieve state of the art results on cauliflower images from ABRC and Arabidopsis images from CVPPP datasets.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Chengzhuan Yang
Summary: Plant leaf identification is a significant challenge in the fields of computer vision and pattern recognition. This article presents a new approach to plant leaf identification, integrating shape and texture characteristics. The proposed method achieves good recognition performance on three benchmark leaf datasets by combining complementary shape and texture features.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Artificial Intelligence
Hao Wu, Lincong Fang, Qian Yu, Jingrong Yuan, Chengzhuan Yang
Summary: This article proposes a novel plant leaf identification method by combining shape and convolutional characteristics. The method achieves good recognition results on nine benchmark leaf datasets for unsupervised leaf image retrieval and supervised leaf image classification, outperforming prior state-of-the-art approaches.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Jameer Gulab Kotwal, Ramgopal Kashyap, Pathan Mohd. Shafi
Summary: This paper proposes a hybrid strategy based on optimized automatic deep learning for plant leaf disease classification. By using techniques such as image preprocessing and disease region segmentation, the accuracy of disease classification is improved. By extracting crucial features and using artificial driving-EfficientNet for classification, a high classification accuracy is achieved.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Viraj K. Gajjar, Anand K. Nambisan, Kurt L. Kosbar
Summary: This study evaluates the performance of EfficientNet and several other CNN architectures on multiple leaf datasets, and constructs an end-to-end CNN classifier using oversampling, undersampling, and transfer learning techniques. The EfficientNet model achieves high accuracy with an F-score on the combined dataset.
Article
Chemistry, Multidisciplinary
Haixia Qi, Yu Liang, Quanchen Ding, Jun Zou
Summary: This study utilized a combination of traditional machine learning methods and deep learning models to identify peanut-leaf diseases, with the deep learning model outperforming traditional methods. After ensemble by logistic regression, ResNet50 and DenseNet121 showed good performance in terms of accuracy and F1 score.
APPLIED SCIENCES-BASEL
(2021)
Article
Computer Science, Artificial Intelligence
S. Nandhini, K. Ashokkumar
Summary: This paper presents an efficient algorithm, MHGSO, to optimize the hyperparameters of the DenseNet-121 architecture. The optimized model achieves higher classification accuracy for different plant images and shows good performance in complex backgrounds. It helps in early detection of plant pathogens and improves agricultural productivity.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Automation & Control Systems
Abdullah Elen, Emre Avuclu
Summary: Plants are essential for oxygen and nutrients on Earth, making conservation of biodiversity crucial for the survival of other species. An image processing-based method has been developed in this study to automatically separate the petiole region of plant leaves, showing promising results in experimental studies.
MEASUREMENT & CONTROL
(2021)
Article
Biochemistry & Molecular Biology
Fatima A. R. Mota, Sarah A. P. Pereira, Andre R. T. S. Araujo, Beatriz Gullon, Marieta L. C. Passos, Maria Lucia M. F. S. Saraiva
Summary: The aim of this study is to develop an automated method for evaluating myeloperoxidase (MPO) activity and test the inhibitory effects of different plant extracts on this enzyme. The study optimized an automatic sequential injection analysis system and demonstrated its accuracy and precision. The study also found that Arbutus unedo L. extract exhibited the highest inhibitory activity on MPO based on its phenolic compound content. The results support the potential use of these plant species for treating infection and inflammation.
Article
Engineering, Civil
M. Pereira, A. M. D'Altri, S. de Miranda, B. Glisic
Summary: This paper proposes an automatic block-by-block pattern generator for multi-leaf nonperiodic masonries, which is then applied in block-based computational analysis of historical masonry structures. The volume of the structure is automatically filled with blocks using the 3D volume of the structure and the 3D block definition of a sample, considering through-thickness blocks and structural details. The efficiency and the generation of statistically-consistent patterns of the filling algorithm are assessed using a meaningful benchmark and applied in full-scale computational analyses.
ENGINEERING STRUCTURES
(2023)
Article
Computer Science, Artificial Intelligence
Xin Li, Siyu Guo, Linrui Gong, Yuan Lan
Summary: This study proposes an automatic detection method for leaf stomatal morphology analysis based on an attention mechanism and deep learning. The method achieves high detection accuracy and mAP on two different datasets.
IET IMAGE PROCESSING
(2023)
Article
Plant Sciences
Yan Zhang, Shiyun Wa, Longxiang Zhang, Chunli Lv
Summary: The detection of plant diseases is crucial in agricultural production. Traditional deep learning algorithms have limitations in this area, and this study proposes a Tranvolution detection network with GAN modules to improve detection performance. The method combines generative models, modified Transformers, and CNNs. Experimental results show satisfactory performance in terms of precision, recall, and mAP.
FRONTIERS IN PLANT SCIENCE
(2022)
Article
Plant Sciences
Ahmed Afifi, Abdulaziz Alhumam, Amira Abdelwahab
Summary: This study developed and evaluated several methods for identifying plant diseases with little data, using convolutional neural networks. The results showed that a baseline model trained with a large set of source field images can be adapted to classify new diseases from a small number of images, outperforming metric learning methods in cases where source domain images significantly differ from target domain images.
Article
Agriculture, Multidisciplinary
Xijian Fan, Peng Luo, Yuen Mu, Rui Zhou, Tardi Tjahjadi, Yi Ren
Summary: This paper proposes a general framework for recognizing plant diseases, which effectively captures and distinguishes the features of plant leaf diseases through feature fusion and center loss.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2022)
Article
Plant Sciences
H. Mattila, D. Valev, K. B. Mishra, V. Havurinne, O. Virtanen, M. Antinluoma, E. Tyystjarvi
Summary: The growth rates during batch and turbidostat modes varied independently of each other, indicating that a growth rate measured in a batch cannot be used to determine the continuous growth rate. Different photoinhibition susceptibilities in tested algae suggest different amounts of energy invested in repair. However, photoinhibition tolerance did not necessarily lead to a fast growth rate at a moderate light intensity.
Article
Multidisciplinary Sciences
Risto Kalliola, Ari Linna, Kalle Ruokolainen, Esa Tyystjarvi, Carl Lange
Summary: In this study, we investigated the element distribution in Guadua bamboo leaves using scanning electron microscopy with energy-dispersive X-ray spectroscopy. The results showed high silicon content and low calcium and potassium content in the leaves. Silicon was mainly found in bulliform cells, bilobate shaped short cells, and stomata. Potassium often surrounded silicon-loaded cells, and silicon and potassium had overlapping distributions in the intercostal areas near vein margins. Calcium showed abundant spotted distribution in the intercostal areas. Moreover, adjacent fusoid cells showed different combinations of silicon, potassium, and calcium, suggesting potentially variable functions of these cells.
SN APPLIED SCIENCES
(2022)
Article
Statistics & Probability
Riikka Numminen, Ileana Montoya Perez, Ivan Jambor, Tapio Pahikkala, Antti Airola
Summary: This article presents a method called quicksort leave-pair-out cross-validation (QLPOCV) to decrease the time complexity of obtaining a reliable ranking of data, aiming to improve the performance evaluation of classifiers. The experimental study demonstrates that QLPOCV produces as accurate performance estimation as the existing method TLPOCV, outperforming other commonly used methods.
COMPUTATIONAL STATISTICS
(2023)
Article
Plant Sciences
Heta Mattila, Sujata Mishra, Taina Tyystjarvi, Esa Tyystjarvi
Summary: Singlet oxygen (O-1(2)) has both harmful and signaling functions in photosynthesis. This study investigated the temperature dependence of O-1(2) production, photoinhibition, and recombination pathways. The results suggest that the miss-associated recombination of P(680)(+)Q(A)(-) is the main source of O-1(2). Furthermore, three parallel photoinhibition mechanisms were identified, with the manganese mechanism dominant in UV radiation and light absorption by Chls mechanism dominating in red light.
Article
Plant Sciences
Olli Virtanen, Esa Tyystjarvi
Summary: This study used HPLC to determine the redox state of the plastoquinone pool (PQ-pool) in Chlamydomonas reinhardtii and compared it with the light state. The results showed that the dynamics of the PQ-pool in C. reinhardtii under different light conditions are more complicated than in plants, possibly due to the larger number of LHC units and less different absorption profiles of the photosystems in C. reinhardtii.
PHOTOSYNTHESIS RESEARCH
(2023)
Article
Plant Sciences
H. Mattila, V. Havurinne, T. Antal, E. Tyystjaervi
Summary: A chlorophyll a fluorescence method was developed to estimate the excitation preference of PSI or PSII in plants under different light wavelengths, which was tested in green microalgae. The results showed variations in the response of different algae to light, possibly due to differences in the redox states.
Article
Plant Sciences
Heta Mattila, Esa Tyystjarvi
Summary: Photosynthetic organisms, such as evergreen plants, may be damaged by strong light at low temperatures. This light-induced damage depends on singlet oxygen and is less pronounced in winter leaves of evergreen plants compared to thylakoids of summer leaves. However, cyanobacteria are not as protected from photoinhibition as evergreen plants, despite their high carotenoid levels.
PHYSIOLOGIA PLANTARUM
(2022)
Article
Plant Sciences
Taras K. Antal, Alena A. Volgusheva, Galina P. Kukarskikh, Evgeniy P. Lukashev, Alexander A. Bulychev, Andrea Margonelli, Silvia Orlanducci, Gabriella Leo, Luciana Cerri, Esa Tyystjarvi, Maya D. Lambreva
Summary: Research has found that single-walled carbon nanotubes (SWCNTs) can protect algal photosynthesis against photoinhibition, and intentional selection of nanomaterial characteristics can overcome their inherent phytotoxicity.
PLANT PHYSIOLOGY AND BIOCHEMISTRY
(2022)
Article
Computer Science, Information Systems
Ilkka Suuronen, Antti Airola, Tapio Pahikkala, Mika Murtojarvi, Valtteri Kaasinen, Henry Railo
Summary: Early detection plays a crucial role in future neuroprotective treatments for Parkinson's disease (PD). Resting state electroencephalographic (EEG) recording has the potential to be a cost-effective method for detecting neurological disorders such as PD. This study explored the impact of the number and placement of electrodes on classifying PD patients and healthy controls using machine learning based on EEG sample entropy. The results indicate that a small subset of electrodes placed in different areas of the brain can achieve classification performance comparable to using a full set of electrodes.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Plant Sciences
Marjaana Rantala, Paula Mulo, Esa Tyystjaervi, Heta Mattila
Summary: During autumn senescence, the disassembly and degradation of photosynthetic protein complexes is poorly understood, as well as the reason why leaves accumulate red pigments. One possibility is that the red pigments, or anthocyanins, protect the senescing leaves from excess light.
PHYSIOLOGIA PLANTARUM
(2023)
Article
Forestry
Heta Mattila, Esa Tyystjarvi
Summary: The reasons behind autumn colors, a striking manifestation of anthocyanin synthesis in plants, are poorly understood. In this study, the researchers investigated the role of red pigments in senescing leaves, and found that they do not provide photoprotection. Instead, the primary function of anthocyanin synthesis appears to be to dispose of carbohydrates, allowing the light reactions to produce energy for nutrient translocation during the last phase of autumn senescence.
Proceedings Paper
Computer Science, Artificial Intelligence
Parisa Movahedi, Valtteri Nieminen, Ileana Montoya Perez, Tapio Pahikkala, Antti Airola
Summary: This study experimentally compares two different protocols for model evaluation and hyperparameter selection for classifiers trained on differentially private synthetic medical data. The results provide novel insights into the practical feasibility and utility of different evaluation protocols.
2023 IEEE 36TH INTERNATIONAL SYMPOSIUM ON COMPUTER-BASED MEDICAL SYSTEMS, CBMS
(2023)