Advancements in Medicinal Plant Identification Using Deep Learning Techniques: A Comprehensive Review

There is a global dependence on medicinal plants for the treatment of diverse health conditions, particularly in developing countries, where government medical facilities are inadequate. Conventional machine learning techniques and traditional methodologies are inadequate for rapid, accurate, and reliable identification and classification of medicinal plants. This has paved the way for deep learning(DL) models to identify and classify medicinal plants. This review focuses on the following questions: 1) Which deep learning models were used for medicinal plant identification and classification and how did they perform? 2) Did hybrid models perform better than ensemble models? 3) What common datasets are used for deep learning-based medicinal plant identification and classification studies? 4) Which data augmentation techniques were prevalent in medicinal plant identification and classification studies? 5) Which performance metrics are frequently used to evaluate medicinal plant identification models? 6) What data splitting ratios are commonly used? 7) Was transfer learning prevalent in medicinal plant identification and classification studies? This review complements the limited literature on medicinal plant identification and classification using deep learning. This study analyzes gaps in the existing literature and provides recommendations for future research. This will enhance the development of robust deep learning models to identify and classify medicinal plants for indigenous knowledge preservation, biodiversity monitoring, and their proper use in pharmaceutical processes.

The accurate identification and monitoring of medicinal plants in remote and diverse terrains is crucial for biodiversity conservation, pharmaceutical research, and traditional medicine documentation. This project presents an IoT-based medicinal plant identification system that integrates deep learning with robotics and wireless communication to automate the process of plant species recognition in real-world environments. Manual plant identification is time-consuming and often error-prone, especially in inaccessible regions. With the advancement of IoT, automated plant classification systems can now be deployed in the field. The ESP32-CAM, a compact low-power module with onboard Wi-Fi and camera, offers an ideal platform for real-time image capture and cloud data transmission. Combining this with the Rocker-Bogie mechanism enables smooth traversal over rough terrain, making it suitable for field applications. The proposed system consists of a mobile robotic platform using the Rocker-Bogie mechanism for stable all-terrain navigation. Mounted on the robot is an ESP32-CAM module that captures high-resolution images of plants from multiple angles. A GPS module is integrated to tag each image with accurate geo location data. The Rocker-Bogie platform provided stable movement across uneven surfaces, while the ESP32-CAM captured clear images for classification. The developed IoT-based medicinal plant identification system effectively combines to provide a reliable and scalable solution for automated plant monitoring in remote environments.

Code smells indicate potential symptoms or problems in software due to inefficient design or incomplete implementation. These problems can affect software quality in the long-term. Code smell detection is fundamental to improving software quality and maintainability, reducing software failure risk, and helping to refactor the code. Previous works have applied several prediction methods for code smell detection. However, many of them show that machine learning (ML) and deep learning (DL) techniques are not always suitable for code smell detection due to the problem of imbalanced data. So, data imbalance is the main challenge for ML and DL techniques in detecting code smells. To overcome these challenges, this study aims to present a method for detecting code smell based on DL algorithms (Bidirectional Long Short-Term Memory (Bi-LSTM) and Gated Recurrent Unit (GRU)) combined with data balancing techniques (random oversampling and Tomek links) to mitigate data imbalance issue. To establish the effectiveness of the proposed models, the experiments were conducted on four code smells datasets (God class, data Class, feature envy, and long method) extracted from 74 open-source systems. We compare and evaluate the performance of the models according to seven different performance measures accuracy, precision, recall, f-measure, Matthew’s correlation coefficient (MCC), the area under a receiver operating characteristic curve (AUC), the area under the precision–recall curve (AUCPR) and mean square error (MSE). After comparing the results obtained by the proposed models on the original and balanced data sets, we found out that the best accuracy of 98% was obtained for the Long method by using both models (Bi-LSTM and GRU) on the original datasets, the best accuracy of 100% was obtained for the long method by using both models (Bi-LSTM and GRU) on the balanced datasets (using random oversampling), and the best accuracy 99% was obtained for the long method by using Bi-LSTM model and 99% was obtained for the data class and Feature envy by using GRU model on the balanced datasets (using Tomek links). The results indicate that the use of data balancing techniques had a positive effect on the predictive accuracy of the models presented. The results show that the proposed models can detect the code smells more accurately and effectively.

Advancements in Raman light sheet microscopy have provided a powerful, non-invasive, marker-free method for imaging complex 3D biological structures, such as cell cultures and spheroids. By combining 3D tomograms made by Rayleigh scattering, Raman scattering, and fluorescence detection, this modality captures complementary spatial and molecular data, critical for biomedical research, histology, and drug discovery. Despite its capabilities, Raman light sheet microscopy faces inherent limitations, including low signal intensity, high noise levels, and restricted spatial resolution, which impede the visualization of fine subcellular structures. Traditional enhancement techniques like Fourier transform filtering and spectral unmixing require extensive preprocessing and often introduce artifacts. More recently, deep learning techniques, which have shown great promise in enhancing image quality, face their own limitations. Specifically, conventional deep learning models require large quantities of high-quality, manually labeled training data for effective denoising and super-resolution tasks, which is challenging to obtain in multi-modal microscopy. In this study, we address these limitations by exploring advanced zero-shot and self-supervised learning approaches, such as ZS-DeconvNet, Noise2Noise, Noise2Void, Deep Image Prior (DIP), and Self2Self, which enhance image quality without the need for labeled and large datasets. This study offers a comparative evaluation of zero-shot and self-supervised learning methods, evaluating their effectiveness in denoising, resolution enhancement, and preserving biological structures in multi-modal Raman light sheet microscopic images. Our results demonstrate significant improvements in image clarity, offering a reliable solution for visualizing complex biological systems. These methods establish the way for future advancements in high-resolution imaging, with broad potential for enhancing biomedical research and discovery.

Comprehensive Study for Breast Cancer Using Deep Learning and Traditional Machine Learning

MediFlora-Net: Quantum-enhanced deep learning for precision medicinal plant identification.

Cancer is considered one of the most aggressive and destructive diseases that shortens the average lives of patients. Misdiagnosed brain tumours lead to false medical intervention, which reduces patients' chance of survival. Accurate early medical diagnoses of brain tumour are an essential point for starting treatment plans that improve the survival of patients with brain tumours. Computer-aided diagnostic systems have provided consecutive successes for helping medical doctors make accurate diagnoses and have conducted positive strides in the field of deep and machine learning. Deep convolutional layers extract strong distinguishing features from the regions of interest compared with those extracted using traditional methods. In this study, different experiments are performed for brain tumour diagnosis by combining deep learning and traditional machine learning techniques. AlexNet and ResNet-18 are used with the support vector machine (SVM) algorithm for brain tumour classification and diagnosis. Brain tumour magnetic resonance imaging (MRI) images are enhanced using the average filter technique. Then, deep learning techniques are applied to extract robust and important deep features via deep convolutional layers. The process of combining deep and machine learning techniques starts, where features are extracted using deep learning techniques, namely, AlexNet and ResNet-18. These features are then classified using SoftMax and SVM. The MRI dataset contains 3,060 images divided into four classes, which are three tumours and one normal. All systems have achieved superior results. Specifically, the AlexNet+SVM hybrid technique exhibits the best performance, with 95.10% accuracy, 95.25% sensitivity, and 98.50% specificity.

Classification and identification of medicinal plants has garnered a lot of attention in the last several years. Many are curious about these plants because of the potential advantages they may have on human health. This study introduces a new deep learning-based artificial intelligence approach to medicinal plant identification using the CNN-X architecture. We are quite proud of our Python-based model's 93% training accuracy and 97% validation accuracy. More than 18,000 images of medical plants organized into 200 distinct categories make up the VNPlant-200 dataset, which our model uses to reliably identify them using a number of visual features. After undergoing extensive training, the CNN-X model can correctly identify a variety of species and detect intricate patterns in the images. The model's performance was greatly enhanced with hyperparameter adjustments and CNN-X architectural modifications. The high accuracies obtained indicate the model's capacity to accurately identify and classify therapeutic plants. This improves the accuracy and efficiency of identification procedures. By making it easier for researchers, botanists, and medical experts to identify therapeutic plants, our AI-based approach advances automated identification medicine. This study demonstrates how deep learning and CNN-X architecture, a kind of artificial intelligence (AI), may enhance the identification of medicinal plants. It expands the number of potential research avenues when applied to the VNPlant-200 dataset, advancing herbal medicine and botanical science. The experiment's overall findings indicate that deep learning techniques can be useful for identifying therapeutic plants, which should spur further developments in the field of herbal medicine.

The pedestrian re-identification problem (i.e., re-id) is essential and pre-requisite in multi-camera video surveillance studies, provided the fact that pedestrian targets need to be accurately re-identified across a network of multiple cameras with non-overlapping fields of views before other post-hoc high-level utilizations (i.e., tracking, behaviors analyses, activities monitoring, etc.) can be carried out. Driven by recent developments in deep learning techniques, the important re-id problem is often tackled via either deep discriminant learning or deep generative learning techniques. However, most contemporary deep learning-based models with tremendously deep structures are not easy to be trained because of the notorious vanishings gradient problem. In this study, a novel full-scaled deep discriminant learning model is proposed. The novelty of the full-scale model is significant, as three crucial concepts in designing a deep learning model, including depth, width, and cardinality, are all taken into consideration, simultaneously. Therefore, the new model needs not to be tremendously deep but is more convenient to be trained. Moreover, based on the new model, a novel deep metric learning method is proposed to further solve the important re-id problem. Technically, two algorithms either based on the conventional SGD (stochastic gradient descent) or an alternative more efficient PGD (proximal gradient descent) are both derived. For experimental analyses, the newly introduced full-scaled deep metric learning method has been comprehensively compared with dozens of popular re-id methods proposed from either deep learning or shallow learning perspectives. Several well-known public re-id datasets have been incorporated and rigorous statistical analyses have been carried out to compare all methods regarding their re-id performance. The superiority of the novel full-scaled deep metric learning method has been substantiated, from the statistical point of view.

Deep Learning (DL) techniques are being used in various critical applications like self-driving cars. DL techniques such as Deep Neural Networks (DNN), Deep Reinforcement Learning (DRL), Federated Learning (FL), and Transfer Learning (TL) are prone to adversarial attacks, which can make the DL techniques perform poorly. Developing such attacks and their countermeasures is the prerequisite for making artificial intelligence techniques robust, secure, and deployable. Previous survey papers only focused on one or two techniques and are outdated. They do not discuss application domains, datasets, and testbeds in detail. There is also a need to discuss the commonalities and differences among DL techniques. In this paper, we comprehensively discussed the attacks and defenses in four popular DL models, including DNN, DRL, FL, and TL. We also highlighted the application domains, datasets, metrics, and testbeds in these fields. One of our key contributions is to discuss the commonalities and differences among these DL techniques. Insights, lessons, and future research directions are also highlighted in detail.

Ardabil is well-known for offering the ideal environment for a good, cheap medicinal herb. Various plant parts are used as essential components in producing natural medicines. According to IUCN (International Union for Conservation of Nature) records, many medicinal plants are on the verge of extinction, so employing image processing and computer vision algorithms to distinguish proof of medicinal plants is critical. As a result, the digitalization of beneficial therapeutic plants is critical for biodiversity preservation. The use of Convolutional Neural Network (CNN)-based techniques to distinguish Indian leaf species is investigated in this research. Several Deep Learning frameworks have recently been used to discern, identify, and characterize various plants. This study is mostly focused on identifying medicinal plants that can be found in rural areas. The Transfer Learning technique selected a well-known pre-trained CNN architecture called mobile net v2. The medical plant dataset was built using 30 different classes of medicinal plants, totaling 3000 photos, and these models were assessed with their pre-trained weights. On a held-out test set, the trained model had an accuracy of 98.05 percent, demonstrating the practicality of this approach.

Medicinal plants have a variety of values and are an important source of new drugs and their lead compounds. They have played an important role in the treatment of cancer, AIDS, COVID-19 and other major and unconquered diseases. However, there are problems such as uneven quality and adulteration. Therefore, it is of great significance to find comprehensive, efficient and modern technology for its identification and evaluation to ensure quality and efficacy. In this study, deep learning, which is superior to conventional identification techniques, was extended to the identification of the part and region of the medicinal plant Paris polyphylla var. yunnanensis from the perspective of spectroscopy. Two pattern recognition models, partial least squares discriminant analysis (PLS-DA) and support vector machine (SVM), were established, and the overall discrimination performance of the three types of models was compared. In addition, we also compared the effects of different sample sizes on the discriminant performance of the models for the first time to explore whether the three models had sample size dependence. The results showed that the deep learning model had absolute superiority in the identification of medicinal plant. It was almost unaffected by factors such as data type and sample size. The overall identification ability was significantly better than the PLS-DA and SVM models. This study verified the superiority of the deep learning from examples, and provided a practical reference for related research on other medicinal plants.

The current development in deep learning is witnessing an exponential transition into automation applications. This automation transition can provide a promising framework for higher performance and lower complexity. This ongoing transition undergoes several rapid changes, resulting in the processing of the data by several studies, while it may lead to time-consuming and costly models. Thus, to address these challenges, several studies have been conducted to investigate deep learning techniques; however, they mostly focused on specific learning approaches, such as supervised deep learning. In addition, these studies did not comprehensively investigate other deep learning techniques, such as deep unsupervised and deep reinforcement learning techniques. Moreover, the majority of these studies neglect to discuss some main methodologies in deep learning, such as transfer learning, federated learning, and online learning. Therefore, motivated by the limitations of the existing studies, this study summarizes the deep learning techniques into supervised, unsupervised, reinforcement, and hybrid learning-based models. In addition to address each category, a brief description of these categories and their models is provided. Some of the critical topics in deep learning, namely, transfer, federated, and online learning models, are explored and discussed in detail. Finally, challenges and future directions are outlined to provide wider outlooks for future researchers.

Data is gathered from millions of patients across various healthcare organizations could be framed as electronic health record (EHR). It comprises of data in various forms that include laboratory results, diagnosis reports, medical images, demographic information about patient, and clinical notes. Transformation in the health care could be done with the help of deep learning techniques since these techniques outperform than the conventional machine learning methods. In addition, a huge and complex dataset gets generated day by day, and thus, it enables the need for the deployment of deep learning models. It is examined that the clinical decision in certain fields that vary from analysis of medical research issues and to suggest as well as prioritize the treatments in order to detect abnormalities, and in addition, the identification of genomic markers in tissue samples is done. Deep learning is applied in EHR system to derive patient representation to provide enhanced clinical predictions, and augments clinical decision systems. Moreover, it is applied in order to detect the disease in the earlier stages, determining the clinical risk, forecasting the future need of regular checkups and in addition prediction of hospitalization in the near future if required. According to statistical report, the deep learning software market is estimated to reach 180 million US dollars in size by 2020. The availability of a huge amount of clinical information particularly EHR has stimulated the growth of deep learning techniques which assist in rapid analysis of patient’s data. Nowadays, EHR is being incorporated with deep learning techniques and tools into EHR systems which provide deep insights for health outcomes. The market for EHR deep learning tools is predicted to exceed $34 billion by the mid of 2020s, motivated massively by an emerging desire to automate tasks and provide deeper insights into clinical issues. The global health IT market is anticipated to be worth a surprising $223.16 billion by 2023, driven in part by deep learning. EHR is a thriving research domain for deep learning researchers since it helps to achieve higher accuracies in medical conditions. Clinical information could be extracted by applying deep learning techniques. The extraction is of various types, namely retrieval of single idea, extracting the association, time-based extraction, and advancement of abbreviation. In EHR representation learning process, concept and patient representations are included to obtain detailed analysis and decisive predictive functionalities. Outcome prediction in deep EHR is categorized as static and temporal prediction to predict patient outcomes. The clinical deep learning is more interpretable under the category of maximum activation, constraints, qualitative clustering, and mimic learning in EHR analysis. The objective of this chapter is to provide an insight into the digital transformation of EHR through deep learning techniques. It discusses the deep learning framework and challenges that occur during the development of deep learning models for EHR. Moreover, it also focuses on the deep learning prediction techniques for various diseases and recent advancements on deep learning techniques in the aspect of providing precise medicine and future generation health care.

Deep learning based physical layer security for terrestrial communications in 5G and beyond networks: A survey