Implementation Of Deep Learning Algorithms Research Articles

Deep‐learning‐powered data analysis in plankton ecology

ABSTRACTThe implementation of deep learning algorithms has brought new perspectives to plankton ecology. Emerging as an alternative approach to established methods, deep learning offers objective schemes to investigate plankton organisms in diverse environments. We provide an overview of deep‐learning‐based methods including detection and classification of phytoplankton and zooplankton images, foraging and swimming behavior analysis, and finally ecological modeling. Deep learning has the potential to speed up the analysis and reduce the human experimental bias, thus enabling data acquisition at relevant temporal and spatial scales with improved reproducibility. We also discuss shortcomings and show how deep learning architectures have evolved to mitigate imprecise readouts. Finally, we suggest opportunities where deep learning is particularly likely to catalyze plankton research. The examples are accompanied by detailed tutorials and code samples that allow readers to apply the methods described in this review to their own data.

FPGA Hardware Acceleration Research and Implementation of Deep Learning Algorithms

The convolutional neural network model is an important algorithm for deep learning, and YOLOv3-tiny based on this model has excellent object detection ability. However, the computational power required by the model is still large, and it is difficult to realize the application in the embedded field. This paper proposes a hardware acceleration method for YOLOv3-tiny and implements it on FPGA platform. Firstly, the fixed-point quantitative processing was carried out for the network, and an appropriate fixed-point strategy was designed with the data accuracy as the index. Secondly, the parallel computing design and pipeline optimization principle were carried out, and the FIFO structure was introduced to shorten the running time. Finally, the experiment was carried out on the Xilinx PYNQ-Z2 platform, and the data were compared with the previous related work.

Real-Time Obstacle Detection for Unmanned Surface Vehicle Maneuver

The rapid advancement and increasing demand for Unmanned Surface Vehicle (USV) technology have drawn considerable attention in various sectors, including commercial, research, and military, particularly in marine and shallow water applications. USVs have the potential to revolutionize monitoring systems in remote areas while reducing labor costs. One critical requirement for USVs is their ability to autonomously integrate Guidance, Navigation, and Control (GNC) technology, enabling self-reliant operation without constant human oversight. However, current study for USV shown the use of traditional method using color detection which is inadequate to detect object with unstable lighting condition. This study addresses the challenge of enabling Autonomous Surface Vehicles (ASVs) to operate with minimal human intervention by enhancing their object detection and classification capabilities. In dynamic environments, such as water surfaces, accurate and rapid object recognition is essential. To achieve this, we focus on the implementation of deep learning algorithms, including the YOLO algorithm, to empower USVs with informed navigation decision-making capabilities. Our research contributes to the field of robotics by designing an affordable USV prototype capable of independent operation characterized by precise object detection and classification. By bridging the gap between advanced visualization techniques and autonomous USV technology, we envision practical applications in remote monitoring and marine operations with object detection. This paper presents the initial phase of our research, emphasizing significance of deep learning algorithms for enhancing USV navigation and decision-making in dynamic environmental conditions, resulting in mAP of 99.51%, IoU of 87.80%, error value of the YOLOv4-tiny image processing algorithm is 0.1542.

Weed Detection in Wheat Crops Using Image Analysis and Artificial Intelligence (AI)

In the present study, we used device visualization in tandem with deep learning to detect weeds in the wheat crop system in actual time. We selected the PMAS Arid Agriculture University research farm and wheat crop fields in diverse weather environments to collect the weed images. Some 6000 images were collected for the study. Throughout the season, tfhe databank was assembled to detect the weeds. For this study, we used two different frameworks, TensorFlow and PyTorch, to apply deep learning algorithms. PyTorch’s implementation of deep learning algorithms performed comparatively better than that of TensorFlow. We concluded that the neural network implemented through the PyTorch framework achieves a superior outcome in speed and accuracy compared to other networks, such as YOLO variants. This work implemented deep learning models for weed detection using different frameworks. While working on real-time detection models, it is very important to consider the inference time and detection accuracy. Therefore, we have compared the results in terms of execution time and prediction accuracy. In particular, the accuracy of weed removal from wheat crops was judged to be 0.89 and 0.91, respectively, with inference times of 9.43 ms and 12.38 ms on the NVIDIA RTX2070 GPU for each picture (640 × 640).

A comparative study on the implementation of deep learning algorithms for detection of hepatic necrosis in toxicity studies

Deep learning has recently become one of the most popular methods of image analysis. In non-clinical studies, several tissue slides are generated to investigate the toxicity of a test compound. These are converted into digital image data using a slide scanner, which is then studied by researchers to investigate abnormalities, and the deep learning method has been started to adopt in this study. However, comparative studies evaluating different deep learning algorithms for analyzing abnormal lesions are scarce. In this study, we applied three algorithms, SSD, Mask R-CNN, and DeepLabV3+, to detect hepatic necrosis in slide images and determine the best deep learning algorithm for analyzing abnormal lesions. We trained each algorithm on 5750 images and 5835 annotations of hepatic necrosis including validation and test, augmented with 500 image tiles of 448 × 448 pixels. Precision, recall, and accuracy were calculated for each algorithm based on the prediction results of 60 test images of 2688 × 2688 pixels. The two segmentation algorithms, DeepLabV3+ and Mask R-CNN, showed over 90% of accuracy (0.94 and 0.92, respectively), whereas SSD, an object detection algorithm, showed lower accuracy. The trained DeepLabV3+ outperformed all others in recall while also successfully separating hepatic necrosis from other features in the test images. It is important to localize and separate the abnormal lesion of interest from other features to investigate it on a slide level. Therefore, we suggest that segmentation algorithms are more appropriate than object detection algorithms for use in the pathological analysis of images in non-clinical studies.

Deep learning algorithms for the early detection of breast cancer: A comparative study with traditional machine learning

Deep learning has been widely applied in breast cancer screening to analyze images obtained from X-rays, ultrasound, magnetic resonances, and biopsies. This study suggests that deep learning can also be used to prescreen for cancer by analyzing heterogeneous data obtained from demographic and anthropometric information of patients, biological markers from routine blood samples, and relative risks from meta-analysis and international databases. In this document, feature selection is applied to a database of 64 women diagnosed with breast cancer and a counterfactual group of 52 healthy women, to identify the best predictors of cancer prescreening. The best predictors are used in k-fold Monte Carlo cross-validation experiments that compare deep learning against machine learning. The results indicate that a deep learning architecture that is fine-tuned using feature selection has the lowest false negative rate (i.e., the lowest Type II errors) and can effectively distinguish between patients with and without cancer. Consequently, deep learning—compared to traditional machine learning—promotes a more accurate detection of malignancies, hence reducing the risk of increased tumor size and cancer spreading to nearby or distant lymph nodes, tissues, or organs, due to late detection. Additionally, compared to machine learning, deep learning has the lowest uncertainty in its predictions, as indicated by the lowest standard deviation of its performance metrics. These findings indicate that deep learning algorithms applied to cancer prescreening offer a radiation-free, non-invasive, and an affordable complement to screening methods based on imagery. The implementation of deep learning algorithms in cancer prescreening helps to identify individuals who may require imaging-based screening, can encourage self-examination, and decreases the psychological drawbacks associated with false positives in cancer screening, ultimately leading to earlier detection of malignancy and reducing the healthcare and societal burden associated with cancer treatment.

Personalized Blood Glucose Prediction for Type 1 Diabetes Using Evidential Deep Learning and Meta-Learning.

The availability of large amounts of data from continuous glucose monitoring (CGM), together with the latest advances in deep learning techniques, have opened the door to a new paradigm of algorithm design for personalized blood glucose (BG) prediction in type 1 diabetes (T1D) with superior performance. However, there are several challenges that prevent the widespread implementation of deep learning algorithms in actual clinical settings, including unclear prediction confidence and limited training data for new T1D subjects. To this end, we propose a novel deep learning framework, Fast-adaptive and Confident Neural Network (FCNN), to meet these clinical challenges. In particular, an attention-based recurrent neural network is used to learn representations from CGM input and forward a weighted sum of hidden states to an evidential output layer, aiming to compute personalized BG predictions with theoretically supported model confidence. The model-agnostic meta-learning is employed to enable fast adaptation for a new T1D subject with limited training data. The proposed framework has been validated on three clinical datasets. In particular, for a dataset including 12 subjects with T1D, FCNN achieved a root mean square error of 18.64±2.60 mg/dL and 31.07±3.62 mg/dL for 30 and 60-minute prediction horizons, respectively, which outperformed all the considered baseline methods with significant improvements. These results indicate that FCNN is a viable and effective approach for predicting BG levels in T1D. The well-trained models can be implemented in smartphone apps to improve glycemic control by enabling proactive actions through real-time glucose alerts.

DSNUNet: An Improved Forest Change Detection Network by Combining Sentinel-1 and Sentinel-2 Images

The use of remote sensing images to detect forest changes is of great significance for forest resource management. With the development and implementation of deep learning algorithms in change detection, a large number of models have been designed to detect changes in multi-phase remote sensing images. Although synthetic aperture radar (SAR) data have strong potential for application in forest change detection tasks, most existing deep learning-based models have been designed for optical imagery. Therefore, to effectively combine optical and SAR data in forest change detection, this paper proposes a double Siamese branch-based change detection network called DSNUNet. DSNUNet uses two sets of feature branches to extract features from dual-phase optical and SAR images and employs shared weights to combine features into groups. In the proposed DSNUNet, different feature extraction branch widths were used to compensate for a difference in the amount of information between optical and SAR images. The proposed DSNUNet was validated by experiments on the manually annotated forest change detection dataset. According to the obtained results, the proposed method outperformed other change detection methods, achieving an F1-score of 76.40%. In addition, different combinations of width between feature extraction branches were analyzed in this study. The results revealed an optimal performance of the model at initial channel numbers of the optical imaging branch and SAR image branch of 32 and 8, respectively. The prediction results demonstrated the effectiveness of the proposed method in accurately predicting forest changes and suppressing cloud interferences to some extent.

Artificial Intelligence-Based Detection of Human Embryo Components for Assisted Reproduction by In Vitro Fertilization.

Assisted reproductive technology is helping humans by addressing infertility using different medical procedures that help in a successful pregnancy. In vitro fertilization (IVF) is one of those assisted reproduction methods in which the sperm and eggs are combined outside the human body in a specialized environment and kept for growth. Assisted reproductive technology is helping humans by addressing infertility using different medical procedures that help in a successful pregnancy. The morphology of the embryological components is highly related to the success of the assisted reproduction procedure. In approximately 3–5 days, the embryo transforms into the blastocyst. To prevent the multiple-birth risk and to increase the chance of pregnancy the embryologist manually analyzes the blastocyst components and selects valuable embryos to transfer to the women’s uterus. The manual microscopic analysis of blastocyst components, such as trophectoderm, zona pellucida, blastocoel, and inner cell mass, is time-consuming and requires keen expertise to select a viable embryo. Artificial intelligence is easing medical procedures by the successful implementation of deep learning algorithms that mimic the medical doctor’s knowledge to provide a better diagnostic procedure that helps in reducing the diagnostic burden. The deep learning-based automatic detection of these blastocyst components can help to analyze the morphological properties to select viable embryos. This research presents a deep learning-based embryo component segmentation network (ECS-Net) that accurately detects trophectoderm, zona pellucida, blastocoel, and inner cell mass for embryological analysis. The proposed method (ECS-Net) is based on a shallow deep segmentation network that uses two separate streams produced by a base convolutional block and a depth-wise separable convolutional block. Both streams are densely concatenated in combination with two dense skip paths to produce powerful features before and after upsampling. The proposed ECS-Net is evaluated on a publicly available microscopic blastocyst image dataset, the experimental segmentation results confirm the efficacy of the proposed method. The proposed ECS-Net is providing a mean Jaccard Index (Mean JI) of 85.93% for embryological analysis.

Implementation of Deep Learning Algorithm on a Custom Dataset for Advanced Driver Assistance Systems Applications

Road hazards such as jaywalking pedestrians, stray animals, unmarked speed bumps, vehicles, and road damage can pose a significant threat in poor visibility conditions. Vehicles are fitted with safety technologies like advanced driver assistance systems (ADAS) and AW (automatic warning) systems to tackle these issues. However, these safety systems are complex and expensive, and these proprietary systems are exclusive to high-end models. The majority of the existing vehicles on the road lacks these systems. The YOLO model (You Only Look Once Architecture) was chosen owing to its lightweight architecture and low inference latency. Since YOLO is an open-source architecture, it can enhance interoperability and feasibility of aftermarket/retrofit ADAS devices, which helps in reducing road fatalities. An ADAS which implements a YOLO-based object detection algorithm to detect and mark obstacles (pedestrians, vehicles, animals, speed breakers, and road damage) using a visual bounding box was proposed. The performance of YOLOv3 and YOLOv5 has been evaluated on the Traffic in the Tamil Nadu Roads dataset. The YOLOv3 model has performed exceptionally well with an F1-Score of 76.3% and an mAP (mean average precision) of 0.755, whereas the YOLOv5 has achieved an F1-Score of 73.7% and an mAP of 0.7263.

A high-performance, hardware-based deep learning system for disease diagnosis.

Modern deep learning schemes have shown human-level performance in the area of medical science. However, the implementation of deep learning algorithms on dedicated hardware remains a challenging task because modern algorithms and neuronal activation functions are generally not hardware-friendly and require a lot of resources. Recently, researchers have come up with some hardware-friendly activation functions that can yield high throughput and high accuracy at the same time. In this context, we propose a hardware-based neural network that can predict the presence of cancer in humans with 98.23% accuracy. This is done by making use of cost-efficient, highly accurate activation functions, Sqish and LogSQNL. Due to its inherently parallel components, the system can classify a given sample in just one clock cycle, i.e., 15.75 nanoseconds. Though this system is dedicated to cancer diagnosis, it can predict the presence of many other diseases such as those of the heart. This is because the system is reconfigurable and can be programmed to classify any sample into one of two classes. The proposed hardware system requires about 983 slice registers, 2,655 slice look-up tables, and only 1.1 kilobits of on-chip memory. The system can predict about 63.5 million cancer samples in a second and can perform about 20 giga-operations per second. The proposed system is about 5–16 times cheaper and at least four times speedier than other dedicated hardware systems using neural networks for classification tasks.

Estimating the Role of Blockchain, Deep Learning and Cryptography algorithms in Cloud Security

Cloud network has become very popular in recent days due to its accessibility merits. The data stored in the cloud environment are accessible by the clients from any location. A reliable shielding approach will protect the data stored in the cloud from the hackers and malwares. Blockchain is one of the recent technologies implemented to the cloud network for storing the location of the saved data in an encrypted ledger format. This saves the stored data location without exploring it to the hacker’s algorithm. Hence the hacking algorithm fails by not knowing the location to be targeted. Deep learning is an advanced technique developed to act like that of the human neurological analysis on several problems. Implementation of deep learning algorithm to the cloud security module identifies the movement of malware and spywares in the cloud storage. Similarly the cryptography is an old technique structured to hide the information with a cover data or cover image. It allows the hacking algorithm to extract only the useless data. This paper reviews the recent advancements in the cloud security with blockchain, deep learning and cryptographic models.

Design and implementation of deep learning algorithm for image discrimination of light field confocal endoscopy for esophageal cancer and gastric cancer

The survival rate of early gastric cancer and esophageal cancer is more than 90%. Confocal endoscopy can detect cell morphology and mucosal glandular structure (depth about 500 μm), presenting a cross-sectional microscopic image unfamiliar to both endoscopists and pathologists. Therefore, using computer-aided diagnosis technology to complete real-time artificial intelligence diagnosis of early esophageal and gastric cancer is of great significance for early detection and early treatment of cancer patients. ResNet convolution neural network based image classification model, the introduction of attention mechanism based on CBAM module to improve the performance of the model, to achieve intelligent diagnosis of confocal endoscopy images.

Implementation of Deep Learning Algorithm and Integrating It with Hand Gesture Recognition

Implementation of Deep Learning Algorithm and Integrating It with Hand Gesture Recognition

Online multiclass EEG feature extraction and recognition using modified convolutional neural network method

<p>Many techniques have been introduced to improve both brain-computer interface (BCI) steps: feature extraction and classification. One of the emerging trends in this field is the implementation of deep learning algorithms. There is a limited number of studies that investigated the application of deep learning techniques in electroencephalography (EEG) feature extraction and classification. This work is intended to apply deep learning for both stages: feature extraction and classification. This paper proposes a modified convolutional neural network (CNN) feature extractorclassifier algorithm to recognize four different EEG motor imagery (MI). In addition, a four-class linear discriminant analysis (LDR) classifier model was built and compared to the proposed CNN model. The paper showed very good results with 92.8% accuracy for one EEG four-class MI set and 85.7% for another set. The results showed that the proposed CNN model outperforms multi-class linear discriminant analysis with an accuracy increase of 28.6% and 17.9% for both MI sets, respectively. Moreover, it has been shown that majority voting for five repetitions introduced an accuracy advantage of 15% and 17.2% for both EEG sets, compared with single trials. This confirms that increasing the number of trials for the same MI gesture improves the recognition accuracy</p>

Skin disease diagnosis with deep learning: A review

Skin cancer is one of the most threatening diseases worldwide. However, diagnosing skin cancer correctly is challenging. Recently, deep learning algorithms have emerged to achieve excellent performance in various tasks. Particularly, they have been applied to the skin disease diagnosis tasks. In this paper, we present a review on deep learning methods and their applications in skin disease diagnosis. We first present a brief introduction to skin diseases and image acquisition methods in dermatology, and list several publicly available skin datasets. Then, we introduce the conception of deep learning, and review popular deep learning architectures and popular frameworks facilitating the implementation of deep learning algorithms. Thereafter, performance evaluation metrics are presented. As an important part of this article, we then review the literature involving deep learning methods for skin disease diagnosis from several aspects according to the specific tasks. Additionally, we discuss the challenges faced in the area and suggest possible future research directions. The major purpose of this article is to provide a conceptual and systematically review of the recent works on skin disease diagnosis with deep learning. Given the popularity of deep learning, there remains great challenges in the area, as well as opportunities that we can explore in the future.

Semantic segmentation of PolSAR image data using advanced deep learning model

Urban area mapping is an important application of remote sensing which aims at both estimation and change in land cover under the urban area. A major challenge being faced while analyzing Synthetic Aperture Radar (SAR) based remote sensing data is that there is a lot of similarity between highly vegetated urban areas and oriented urban targets with that of actual vegetation. This similarity between some urban areas and vegetation leads to misclassification of the urban area into forest cover. The present work is a precursor study for the dual-frequency L and S-band NASA-ISRO Synthetic Aperture Radar (NISAR) mission and aims at minimizing the misclassification of such highly vegetated and oriented urban targets into vegetation class with the help of deep learning. In this study, three machine learning algorithms Random Forest (RF), K-Nearest Neighbour (KNN), and Support Vector Machine (SVM) have been implemented along with a deep learning model DeepLabv3+ for semantic segmentation of Polarimetric SAR (PolSAR) data. It is a general perception that a large dataset is required for the successful implementation of any deep learning model but in the field of SAR based remote sensing, a major issue is the unavailability of a large benchmark labeled dataset for the implementation of deep learning algorithms from scratch. In current work, it has been shown that a pre-trained deep learning model DeepLabv3+ outperforms the machine learning algorithms for land use and land cover (LULC) classification task even with a small dataset using transfer learning. The highest pixel accuracy of 87.78% and overall pixel accuracy of 85.65% have been achieved with DeepLabv3+ and Random Forest performs best among the machine learning algorithms with overall pixel accuracy of 77.91% while SVM and KNN trail with an overall accuracy of 77.01% and 76.47% respectively. The highest precision of 0.9228 is recorded for the urban class for semantic segmentation task with DeepLabv3+ while machine learning algorithms SVM and RF gave comparable results with a precision of 0.8977 and 0.8958 respectively.

A novel hybrid model for intrusion detection systems in SDNs based on CNN and a new regularization technique

Software-defined networking (SDN) is a new networking paradigm that separates the controller from the network devices i.e. routers and switches. The centralized architecture of the SDN facilitates the overall network management and addresses the requirement of current data centers. While there are high benefits offered by the SDN architecture, the risk of new attacks is a critical problem and can prevent the wide adoption of SDNs. The SDN controller is a crucial element, and it is an attractive target for the intruders. In case the attacker successfully accessed the SDN controller, it can route the traffic based on its own requirements, causing severe damage to the entire network. The network intrusion detection systems (NIDSs) are important tools to detect and secure the network environment from malicious activities and anomalous attacks. Deep Learning (DL) has recently shown desirable results in a variety of problems, such as text, speech, and image applications, etc.While several related works deployed DL for NIDSs, most of these approaches ignore the influence of the overfitting problem during the implementation of DL algorithms. As a result, it can impact the robustness of the anomaly detection system and lead to poor model performance for zero-day attacks. In this work, we propose a new hybrid DL approach based on the convolutional neural network (CNN) to classify the flow traffic into normal or attack classes. A new regularizer method, namely SD-Reg, which is based on the standard deviation of the weight matrix, has been used to address the problem of overfitting and to improve the capability of NIDSs in detection of unseen intrusion events. The evaluation results indicate that the SD-Reg outperforms the previous regularizer methods. In addition, the proposed hybrid technique gives a higher performance in all the evaluation metrics compared to the single DL models. Several datasets, including the InSDN – the most recent dataset for SDN – are used to train and evaluate the performance of all techniques. Furthermore, we suggest a lightweight NIDS by training the CNN-based models using a less number of features without causing a significant drop in the model performance.

Evaluation of ASR in concrete using acoustic emission and deep learning

Alkali-silica reaction (ASR) is one of main damages causes in concrete structures such as nuclear power plants which may endanger structural serviceability and integrity. Acoustic emission (AE) is a passive nondestructive method for structural health monitoring. It is very sensitive and has the capability of monitoring structures continuously. This method may be an alternative for early damage detection in concrete nuclear structures affected by ASR. The innovation of this paper lies in the implementation of deep learning algorithms to evaluate the ASR progress. ASR was monitored by AE in a concrete specimen, which was cast with reactive coarse aggregates and reinforced by steel rebars. The AE signals recorded during the experiment were filtered and divided into two classes. Two deep learning algorithms of convolutional neural network (CNN) and stacked autoencoder were employed to classify the AE signals into the corresponding classes. The model based on CNN resulted in a classifier with higher accuracy than the model based on the autoencoder network.

Transductive Transfer Learning for Domain Adaptation in Brain Magnetic Resonance Image Segmentation.

Segmentation of brain images from Magnetic Resonance Images (MRI) is an indispensable step in clinical practice. Morphological changes of sub-cortical brain structures and quantification of brain lesions are considered biomarkers of neurological and neurodegenerative disorders and used for diagnosis, treatment planning, and monitoring disease progression. In recent years, deep learning methods showed an outstanding performance in medical image segmentation. However, these methods suffer from generalisability problem due to inter-centre and inter-scanner variabilities of the MRI images. The main objective of the study is to develop an automated deep learning segmentation approach that is accurate and robust to the variabilities in scanner and acquisition protocols. In this paper, we propose a transductive transfer learning approach for domain adaptation to reduce the domain-shift effect in brain MRI segmentation. The transductive scenario assumes that there are sets of images from two different domains: (1) source—images with manually annotated labels; and (2) target—images without expert annotations. Then, the network is jointly optimised integrating both source and target images into the transductive training process to segment the regions of interest and to minimise the domain-shift effect. We proposed to use a histogram loss in the feature level to carry out the latter optimisation problem. In order to demonstrate the benefit of the proposed approach, the method has been tested in two different brain MRI image segmentation problems using multi-centre and multi-scanner databases for: (1) sub-cortical brain structure segmentation; and (2) white matter hyperintensities segmentation. The experiments showed that the segmentation performance of a pre-trained model could be significantly improved by up to 10%. For the first segmentation problem it was possible to achieve a maximum improvement from 0.680 to 0.799 in average Dice Similarity Coefficient (DSC) metric and for the second problem the average DSC improved from 0.504 to 0.602. Moreover, the improvements after domain adaptation were on par or showed better performance compared to the commonly used traditional unsupervised segmentation methods (FIRST and LST), also achieving faster execution time. Taking this into account, this work presents one more step toward the practical implementation of deep learning algorithms into the clinical routine.