Deep Learning Research Research Articles

A Comparative Review on Enhancing Visual Simultaneous Localization and Mapping with Deep Semantic Segmentation.

Visual simultaneous localization and mapping (VSLAM) enhances the navigation of autonomous agents in unfamiliar environments by progressively constructing maps and estimating poses. However, conventional VSLAM pipelines often exhibited degraded performance in dynamic environments featuring mobile objects. Recent research in deep learning led to notable progress in semantic segmentation, which involves assigning semantic labels to image pixels. The integration of semantic segmentation into VSLAM can effectively differentiate between static and dynamic elements in intricate scenes. This paper provided a comprehensive comparative review on leveraging semantic segmentation to improve major components of VSLAM, including visual odometry, loop closure detection, and environmental mapping. Key principles and methods for both traditional VSLAM and deep semantic segmentation were introduced. This paper presented an overview and comparative analysis of the technical implementations of semantic integration across various modules of the VSLAM pipeline. Furthermore, it examined the features and potential use cases associated with the fusion of VSLAM and semantics. It was found that the existing VSLAM model continued to face challenges related to computational complexity. Promising future research directions were identified, including efficient model design, multimodal fusion, online adaptation, dynamic scene reconstruction, and end-to-end joint optimization. This review shed light on the emerging paradigm of semantic VSLAM and how deep learning-enabled semantic reasoning could unlock new capabilities for autonomous intelligent systems to operate reliably in the real world.

Tomato Leaf Disease Detection using Flask Frame Work

In recent years, plant leaf diseases has become a widespread problem for which an accurate research and rapid application of deep learning in plant disease classification is required, tomato is also one of the most important plants and seeds which are used worldwide for cooking in either dried or fresh form, tomato are a great source of protein that offer many health benefits, but there are a lot of diseases associated with tomato leaf which hinder its production. Thus, an accurate classification of tomato leaf diseases is needed to solve the problem in the early stage. A deep learning approach is proposed to identify and classify leaf disease by using public dataset of leaf image and CNN model with the open source library TensorFlow. In this project, we proposed a method to classify tomato leaf disease and to find and describe the efficient network architecture (hyper parameters and optimization methods). Moreover, after applying each architecture separately, we compared their obtained results to find out the best architecture configuration for classifying tomato leaf diseases and their results. Furthermore, to satisfy the classification requirements, the model was trained using CNN architecture check if we could get faster training times, higher accuracy and easier retraining. Deep learning is a branch of artificial intelligence. In recent years, with the advantages of automatic learning and feature extraction, it has been widely concerned by academic and industrial circles. It has been widely used in image and video processing, voice processing, and natural language processing. At the same time, it has also become a research hotspot in the field of agricultural plant protection, such as plant disease recognition. The application of deep learning in plant disease recognition can avoid the disadvantages caused by artificial selection of disease spot features, make plant disease feature extraction more objective, and improve the research efficiency and technology transformation speed. This review provides the research progress of deep learning technology in the field of crop leaf disease identification in recent years. In this project, we present the current trends and challenges for the detection of plant leaf disease using deep learning and advanced imaging techniques. We hope that this project will be a valuable resource for researchers who study the detection of plant diseases. At the same time, we also discussed some of the current challenges and problems that need to be resolved.

Improving the Generalizability and Performance of an Ultrasound Deep Learning Model Using Limited Multicenter Data for Lung Sliding Artifact Identification.

Deep learning (DL) models for medical image classification frequently struggle to generalize to data from outside institutions. Additional clinical data are also rarely collected to comprehensively assess and understand model performance amongst subgroups. Following the development of a single-center model to identify the lung sliding artifact on lung ultrasound (LUS), we pursued a validation strategy using external LUS data. As annotated LUS data are relatively scarce-compared to other medical imaging data-we adopted a novel technique to optimize the use of limited external data to improve model generalizability. Externally acquired LUS data from three tertiary care centers, totaling 641 clips from 238 patients, were used to assess the baseline generalizability of our lung sliding model. We then employed our novel Threshold-Aware Accumulative Fine-Tuning (TAAFT) method to fine-tune the baseline model and determine the minimum amount of data required to achieve predefined performance goals. A subgroup analysis was also performed and Grad-CAM++ explanations were examined. The final model was fine-tuned on one-third of the external dataset to achieve 0.917 sensitivity, 0.817 specificity, and 0.920 area under the receiver operator characteristic curve (AUC) on the external validation dataset, exceeding our predefined performance goals. Subgroup analyses identified LUS characteristics that most greatly challenged the model's performance. Grad-CAM++ saliency maps highlighted clinically relevant regions on M-mode images. We report a multicenter study that exploits limited available external data to improve the generalizability and performance of our lung sliding model while identifying poorly performing subgroups to inform future iterative improvements. This approach may contribute to efficiencies for DL researchers working with smaller quantities of external validation data.

Online sorting of surface defective walnuts based on deep learning

To address concerns regarding walnut shell damage and inadequate sorting precision during the mechanized sorting of walnuts, a walnut automatic sorting machine was designed based on deep learning and experimental research. Initially, the rationality of the design was verified through experiment. Then, three deep learning semantic segmentation algorithms, namely PSPnet, U-net, and Deeplabv3+, were selected to train walnut detection models. Results indicated that the U-net algorithm proved to be the most effective, achieving a Mean Intersection over Union of 96.71% and a Mean Pixel Accuracy value of 98.52%. Finally, performance tests were conducted on the prototype machine, yielding results with an average sorting efficiency of 51.70 kg/h, an average loss rate of 6.50%, and an average accuracy of sorting walnuts of 92.98%. The findings can provide insights for future structural improvements and operational parameter optimization of walnut automatic sorting machines.

Advancements in Crop Disease Detection: Analytical Methods for Recognizing Disease Stages through Leaf Analysis

Plant diseases have the potential to damage the livelihoods of farmers and impede their capacity to generate an adequate quantity of food. Diagnosis and early detection of plant diseases are critical for their effective management and control. Leaf analysis shows great potential as a method for forecasting disease stages due to its ability to detect subtle alterations in leaf physiology and appearance that may occur prior to the onset of conspicuous symptoms. By utilizing the algorithms of machine learning and deep learning, it is possible to classify characteristics extracted from photographs of leaves into distinct disease stages. Recent studies have demonstrated the potential of these algorithms, as they have achieved remarkable accuracy in disease stage prediction despite having limited training data. The integration of machine learning and deep learning techniques with foliage analysis holds promise for revolutionizing plant disease management through the provision of timely identification, accurate diagnosis, and customized treatment. In order to formulate efficacious disease management strategies, precise determination of the developmental stage of plant diseases is imperative. Scholars are presently devising novel approaches to identify the stages of plant diseases by employing diverse methodologies, including spectroscopy, machine learning, and image processing. This can benefit producers in substantial economic and environmental ways, as well as contribute to the improvement of food security. The primary investigation comprises an assortment of articles spanning the years 2014 to 2023. After evaluating various search strategies, a total of 117 research publications were identified, of which 43 were pertinent. The article examines numerous developments in deep learning research. In addition, it will facilitate the assessment of the present and prospective state of plant disease research by employing deep learning methodologies.

3D Point Cloud Dataset of Heavy Construction Equipment

Object recognition algorithms and datasets based on point cloud data have been mainly designed for autonomous vehicles. When applied to the construction industry, they face challenges due to the origin of point cloud data from large earthwork sites, resulting in high volumes of data and density. This research prioritized the development of 3D point cloud datasets specifically for heavy construction equipment, including dump trucks, rollers, graders, excavators, and dozers; all of which are extensively used in earthwork sites. The aim was to enhance the efficiency and productivity of machine learning (ML) and deep learning (DL) research that relies on 3D point cloud data in the construction industry. Notably, unlike conventional approaches to acquiring point cloud data using UAVs (Unmanned Aerial Vehicles) and UGVs (Unmanned Ground Vehicles), the datasets for the five types of heavy construction equipment established in this research were generated using 3D-scanned diecast models of heavy construction equipment to create point cloud data.

COMPARATIVE ANALYSIS ON DEEP LEARNING METHODS FOR BOT DETECTION

Social bots are automated accounts on social media that are managed by software and overseen by people behind the scenes. While some bots are created for positive reasons, like sharing news updates or offering assistance in crises, others have been misused to spread false information, rumors, or influence political events. There are tools in place to identify and eliminate harmful bots automatically, but creators are continually evolving their techniques to evade detection. Hence, there is a pressing need for improved ways to differentiate between genuine and automated bot accounts. In recent years, several research studies have delved into the realm of social media bot detection, offering an extensive overview of different detection strategies, including advanced approaches such as machine learning (ML) and deep learning (DL) techniques. To the best of our knowledge, this study stands out as the first to solely focus on DL techniques, evaluating their efficacy and rationale in comparison to each other and to conventional ML methods. In this research, we offer a comprehensive classification of the characteristics utilized in deep learning research and provide information on the necessary preprocessing methods for creating appropriate training data for a deep learning model. We highlight the gaps identified in review papers discussing deep learning and machine learning studies, offering insights into future developments in this area. In general, deep learning methods have proven to be efficient in terms of computation and time for detecting social bots, demonstrating superior or comparable performance compared to traditional machine learning techniques. Key Words: Bot Detection, Social Media, Deep Learning, Computation, Automation

Characterizing Deep Learning Package Supply Chains in PyPI: Domains, Clusters, and Disengagement

Deep learning (DL) frameworks have become the cornerstone of the rapidly developing DL field. Through installation dependencies specified in the distribution metadata, numerous packages directly or transitively depend on DL frameworks, layer after layer, forming DL package supply chains (SCs), which are critical for DL frameworks to remain competitive. However, vital knowledge on how to nurture and sustain DL package SCs is still lacking. Achieving this knowledge may help DL frameworks formulate effective measures to strengthen their SCs to remain competitive and shed light on dependency issues and practices in the DL SC for researchers and practitioners. In this paper, we explore the domains, clusters, and disengagement of packages in two representative PyPI DL package SCs to bridge this knowledge gap. We analyze the metadata of nearly six million PyPI package distributions and construct version-sensitive SCs for two popular DL frameworks: TensorFlow and PyTorch. We find that popular packages (measured by the number of monthly downloads) in the two SCs cover 34 domains belonging to eight categories. Applications , Infrastructure , and Sciences categories account for over 85% of popular packages in either SC and TensorFlow and PyTorch SC have developed specializations on Infrastructure and Applications packages, respectively. We employ the Leiden community detection algorithm and detect 131 and 100 clusters in the two SCs. The clusters mainly exhibit four shapes: Arrow, Star, Tree, and Forest with increasing dependency complexity. Most clusters are Arrow or Star, while Tree and Forest clusters account for most packages (Tensorflow SC: 70.7%, PyTorch SC: 92.9%). We identify three groups of reasons why packages disengage from the SC (i.e., remove the DL framework and its dependents from their installation dependencies): dependency issues, functional improvements, and ease of installation. The most common reason in TensorFlow SC is dependency incompatibility and in PyTorch SC is to simplify functionalities and reduce installation size. Our study provides rich implications for DL framework vendors, researchers, and practitioners on the maintenance and dependency management practices of PyPI DL SCs.

A novel technique predicting the rice leaf diseases using Convolutional Neural Network

Various ailments affect rice, a staple crop in India, across different stages of its growth. Identification of these diseases manually poses a significant challenge, especially for farmers lacking in-depth knowledge. Recently, there's been promising advancement in deep learning research through automated picture identification systems employing Convolutional Neural Network (CNN) models. To tackle the scarcity of rice leaf disease image datasets, we developed a deep learning model using Transfer Learning on a limited dataset. Our approach leverages VGG-16 to train and evaluate the proposed CNN architecture, drawing from rice field and internet datasets. Impressively, the model achieves a 95 percent accuracy rate. Key terms in this study include Deep Learning, Convolutional Neural Network (CNN), fine-tuning, and rice leaf diseases.

Lob-based deep learning models for stock price trend prediction: a benchmark study

The recent advancements in Deep Learning (DL) research have notably influenced the finance sector. We examine the robustness and generalizability of fifteen state-of-the-art DL models focusing on Stock Price Trend Prediction (SPTP) based on Limit Order Book (LOB) data. To carry out this study, we developed LOBCAST, an open-source framework that incorporates data preprocessing, DL model training, evaluation, and profit analysis. Our extensive experiments reveal that all models exhibit a significant performance drop when exposed to new data, thereby raising questions about their real-world market applicability. Our work serves as a benchmark, illuminating the potential and the limitations of current approaches and providing insight for innovative solutions.

Research and Application of Deep Learning in Medical Image Reconstruction and Enhancement

In recent years, deep learning technology has made remarkable progress in medical image reconstruction and enhancement, and has become one of the research hotspots in the field of medical image processing. This paper discusses the latest research progress and application of deep learning in medical image reconstruction and enhancement. Firstly, the importance of medical image reconstruction and enhancement and the limitations of traditional methods are introduced. Then, a detailed discussion was conducted on the application of deep learning models, including Convolutional Neural Networks (CNN), Generative Adversarial Networks (GAN), and Autoencoders, in medical image processing. Specifically, an analysis and comparison were conducted on the image reconstruction ability of CNN models, the image enhancement effect of GAN models, and the image denoising and reconstruction of Autoencoder models. Then, the advantages and challenges of deep learning model in medical image processing are discussed, and the future development direction is discussed. Finally, the research results of this paper are summarized and the prospect of future research is put forward. The research in this paper provides some enlightenment and reference for researchers and practitioners in the field of medical image processing, which is helpful to promote the continuous innovation and progress of medical image processing technology.

GCARDTI: Drug–target interaction prediction based on a hybrid mechanism in drug SELFIES

AbstractThe prediction of the interaction between a drug and a target is the most critical issue in the fields of drug development and repurposing. However, there are still two challenges in current deep learning research: (i) the structural information of drug molecules is not fully explored in most drug target studies, and the previous drug SMILES does not correspond well to effective drug molecules and (ii) exploration of the potential relationship between drugs and targets is in need of improvement. In this work, we use a new and better representation of the effective molecular graph structure, SELFIES. We propose a hybrid mechanism framework based on convolutional neural network and graph attention network to capture multi‐view feature information of drug and target molecular structures, and we aim to enhance the ability to capture interaction sites between a drug and a target. In this study, our experiments using two different datasets show that the GCARDTI model outperforms a variety of different model algorithms on different metrics. We also demonstrate the accuracy of our model through two case studies.

Programming Language Selection for The Development of Deep Learning Library

Recently, deep learning has become very successful in various applications, leading to an increasing need for software tools to keep up with the rapid pace of innovation in deep learning research. As a result, we suggested the development of a software library related to deep learning that would be useful for researchers and practitioners in academia and industry for their research endeavors. The programming language is the core of deep learning library development, so this paper describes the selection stage to find the most suitable programming language for developing a deep learning library based on two criteria, including coverage on many projects and the ability to handle high-dimensional array processing. We addressed the comparison of programming languages with two approaches. First, we looked for the most demanding programming languages for AI Jobs by conducting a data-driven approach against the data gathered from several Job-Hunting Platforms. Then, we found the findings that imply Python, C++, and Java as the top three. After that, we compared the three most widely used programming languages by calculating interval time to three different programs that contain an array of exploitation processes. Based on the result of the experiments that were executed in the computer terminal, Java outperformed Python and C++ in two of the three experiments conducted with 5,4047 milliseconds faster than C++ and 231,1639 milliseconds faster than Python to run quick sort algorithm for arrays that contain 100.000 integer values.

Flow prediction of heterogeneous nanoporous media based on physical information neural network

The simulation and prediction of fluid flow in porous media play a profoundly significant role in today's scientific and engineering domains, particularly in gaining a deeper understanding of phenomena such as the migration and fluid flow in underground rock formations and the enhancement of oil recovery rates. The flow of fluids in nanoscale porous media requires consideration of the effects of microscale phenomena, which are challenging to accurately describe using traditional physical models. Currently, research in deep learning for porous media predominantly focuses on conventional porous media, and there is an urgent need for investigations into heterogeneous nanoporous media. Simultaneously, there is a necessity to overcome the limitations of traditional data-driven models lacking physical prior knowledge. Therefore, the integration of physics-informed neural networks, which combine deep learning with physical principles, becomes essential for inferring relatively accurate results from sparse data. In this work, based on the heterogeneity of porous media in shale, we have introduced a deep learning model that couples physical information to predict the flow in heterogeneous nanoscale porous media. In the Physical Information Neural Network model, we utilize point clouds and couple them with deep residual networks. Discrete sampling points are used as inputs, and a multi-level residual connection, along with dimension concatenation, is employed to fuse feature information. The network, through backpropagation, takes into account the Navier-Stokes equations and wall conditions in heterogeneous nanoscale porous media. The results indicate that the apparent permeability and pressure field accuracy are over 90% and 95%, respectively. The Physical Information Neural Network demonstrates promising prospects for predicting flow in nanoscale porous media. Future work will extend to the multiphase complex flow in three-dimensional porous media.

RNA3DB: A structurally-dissimilar dataset split for training and benchmarking deep learning models for RNA structure prediction

With advances in protein structure prediction thanks to deep learning models like AlphaFold, RNA structure prediction has recently received increased attention from deep learning researchers. RNAs introduce substantial challenges due to the sparser availability and lower structural diversity of the experimentally resolved RNA structures in comparison to protein structures. These challenges are often poorly addressed by the existing literature, many of which report inflated performance due to using training and testing sets with significant structural overlap. Further, the most recent Critical Assessment of Structure Prediction (CASP15) has shown that deep learning models for RNA structure are currently outperformed by traditional methods.In this paper we present RNA3DB, a dataset of structured RNAs, derived from the Protein Data Bank (PDB), that is designed for training and benchmarking deep learning models. The RNA3DB method arranges the RNA 3D chains into distinct groups (Components) that are non-redundant both with regard to sequence as well as structure, providing a robust way of dividing training, validation, and testing sets. Any split of these structurally-dissimilar Components are guaranteed to produce test and validations sets that are distinct by sequence and structure from those in the training set. We provide the RNA3DB dataset, a particular train/test split of the RNA3DB Components (in an approximate 70/30 ratio) that will be updated periodically. We also provide the RNA3DB methodology along with the source-code, with the goal of creating a reproducible and customizable tool for producing structurally-dissimilar dataset splits for structural RNAs.

Style Adaptation module: Enhancing detector robustness to inter-manufacturer variability in surface defect detection

In recent years, deep learning-based approaches for industrial surface defect detection have shown great promise. To address the domain shift issue among data from different sources in the industrial domain, we present a novel plug-and-play Style Adaptation (SA) module, which endows the equipped defect detector with the capability to exhibit robustness to diverse styles present within the samples. This module effectively leverages datasets sourced from diverse origins while possessing congruent data types. In contrast to other domain adaptation approaches lacking well-defined domain delineations, the SA module generates representations characterized by distinct practical implications and precise mathematical formulations. Moreover, incorporating attention mechanisms reduces the need for manual intervention, allowing the module to focus autonomously on crucial branches in it. Experimental results demonstrate the superior efficacy of our approach compared to state-of-the-art techniques. Furthermore, an authentic dataset from various manufacturers is publicly available for deep learning research and industrial applications. Access the dataset at: https://github.com/THU-PMVAI/MTS3D

Decoding phenotypic screening: A comparative analysis of image representations

Biomedical imaging techniques such as high content screening (HCS) are valuable for drug discovery, but high costs limit their use to pharmaceutical companies. To address this issue, The JUMP-CP consortium released a massive open image dataset of chemical and genetic perturbations, providing a valuable resource for deep learning research. In this work, we aim to utilize the JUMP-CP dataset to develop a universal representation model for HCS data, mainly data generated using U2OS cells and CellPainting protocol, using supervised and self-supervised learning approaches. We propose an evaluation protocol that assesses their performance on mode of action and property prediction tasks using a popular phenotypic screening dataset. Results show that the self-supervised approach that uses data from multiple consortium partners provides representation that is more robust to batch effects whilst simultaneously achieving performance on par with standard approaches. Together with other conclusions, it provides recommendations on the training strategy of a representation model for HCS images.

BRIDGING PANDEMICS AND PIXELS: A COMPREHENSIVE BIBLIOMETRIC ANALYSIS OF DEEP LEARNING APPLICATIONS IN COVID-19 DETECTION

The world has been significantly impacted by the global pandemic of COVID-19, leading researchers to explore various methods for detecting the virus. Deep learning (DL) technologies have emerged as pivotal tools in detecting and managing the virus. This review article aims to provide a comprehensive examination of the advancements and applications of DL in the context of COVID-19 detection, offering insights into the evolution, impact, and future direction of this rapidly evolving field. A thorough bibliometric review was carried out using the Scopus database. The methodology involved keyword-based searches, analyses of publication trends, and studies of co-citation networks, with a focus on literature from 2020 to 2023. Data visualisation tools, particularly VOSviewer, were used to analyse and map bibliometric data, highlighting publication trends, and authorship patterns. The study found a substantial rise in DL research for detecting COVID-19 from 2020 to 2023, with significant input from countries such as India, China, and Saudi Arabia. The trends in research are expected to showcase developments in DL models for medical imaging, including CT scans and X-rays, as well as the increasing significance of AI in medical diagnostics. The study also pinpointed the main collaborative networks and popular keywords in this research area. This bibliometric analysis aims to lay the foundation for future research directions by offering a comprehensive overview of the evolution of DL in COVID-19 detection It will provide strategic insights and research directions to advance this vital domain at the intersection of computational intelligence and global health.

Modified Structure of Deep Neural Network for Training Multi-Fidelity Data With Non-Common Input Variables

Abstract Multi-fidelity surrogate (MFS) modeling technology, which efficiently constructs surrogate models using low-fidelity (LF) and high-fidelity (HF) data, has been studied to enhance the predictive capability of engineering performances. In addition, several neural network (NN) structures for MFS modeling have been introduced, benefiting from recent developments in deep learning research. However, existing multi-fidelity (MF) NNs have been developed assuming identical sets of input variables for LF and HF data, a condition that is often not met in practical engineering systems. Therefore, this study proposes a new structure of composite NN designed for MF data with different input variables. The proposed network structure includes an input mapping network that connects the LF and HF data's input variables. Even when the physical relationship between these variables is unknown, the input mapping network can be concurrently trained during the process of training the whole network model. Customized loss functions and activation variables are suggested in this study to facilitate forward and backward propagation for the proposed NN structures when training MF data with different inputs. The effectiveness of the proposed method, in terms of prediction accuracy, is demonstrated through mathematical examples and practical engineering problems related to tire performances. The results confirm that the proposed method offers better accuracy than existing surrogate models in most problems. Moreover, the proposed method proves advantageous for surrogate modeling of nonlinear or discrete functions, a characteristic feature of NN-based methods.

Skin cancer classification using Inception Network

Since skin disease is a universally recognized condition among humans, there has been a growing interest in utilizing intelligent systems to classify various skin ailments. This line of research in deep learning holds immense significance for dermatologists. However, accurately determining the presence of a disease is a formidable task due to the intricate nature of skin texture and the visual similarities between different diseases. To address this challenge, skin images undergo filtration to eliminate unwanted noise and undergo further processing to enhance the overall quality of the image. The primary purpose of this study is to construct a deep neural network-based model that is capable of automatically classifying several types of skin cancer as either melanoma or non-melanoma with a prominent level of accuracy. We propose an optimized Inception architecture, in which the InceptionNet model is enhanced with data augmentation and basic layers. The strategy that has been proposed enhances the model's capacity to deal with incomplete and inconsistent data. A dataset of 2637 skin images are used to demonstrate the benefits of the technique that has been proposed. We analyze the performance of the suggested method by looking at its precision, sensitivity, specificity, F1-score, and area under the ROC curve. Proposed InceptionNet provides an accuracy of 84.39% and 85.94%, respectively for Adam and Nadam optimizer. The training process in each subsequent layer exhibits a notable enhancement in effectiveness. An examination of this inquiry can assist experts in making early diagnoses, thereby providing them with insight into the nature of the infection and enabling them to initiate the necessary treatment, if deemed necessary.