Efficient Convolutional Neural Network Research Articles

DTS: dynamic training slimming with feature sparsity for efficient convolutional neural network

DTS: dynamic training slimming with feature sparsity for efficient convolutional neural network

CNNs for remote extraction of urban features: A survey-driven benchmarking

Accurate extraction of urban features such as buildings and roads lays the foundation for the current trends of digital twins of urban systems to support planning, monitoring, navigation, and decision processes. The process of such extraction involves training convolutional neural networks (CNNs) on high-resolution earth observation (EO) images. The spatial resolution of images has increased to a centimetre level and the CNNs are fast evolving in computer vision. The last 10 years of this development have resulted in both high-performance and computationally efficient CNNs, but they are merely benchmarked under a uniform setting. We present a survey-driven benchmark of CNNs starting with a systematic survey of 165 research articles to understand the state-of-the-art of urban feature extraction. The survey looks for the most prominent urban feature, EO source, benchmark dataset, CNN-based deep learning configuration, and hyperparameters. Further, more CNNs are searched in the computer vision domain. Identified from the survey and search, 65 CNNs are trained and evaluated in an encoder–decoder configuration using a benchmark dataset under uniform settings. Extensive hyperparameter tuning of the best-performing CNN is performed with six optimisers and nine loss functions. The tuned CNN is then tested as an encoder in other state-of-the-art encoder–decoder networks. The CNNs and network configurations with the highest scores are further benchmarked on the Massachusetts Building and WHU Building datasets. The findings from this survey-driven benchmark of CNNs will be useful for both academia and industry involved in the science of earth observation and computer vision.

Handwritten Digit Recognition

Abstract: This paper presents a powerful Handwritten Digit Recognition System that combines a Graphical User Interface (GUI) based on Tkinter with Convolutional Neural Networks (CNNs). Our approach, which makes use of the MNIST dataset, includes careful data pretreatment to facilitate efficient CNN model training. Convolutional and pooling layers are included in the design of the model, and they are optimized with the Adam optimizer for higher learning rates. The evaluation's findings demonstrate excellent memory, accuracy, and precision. Furthermore, real-time digit drawing is made possible via an intuitive Tkinter GUI, which confirms the model's applicability. By showing the effectiveness of CNNs and offering an interactive platform for natural user interaction, the research provides a holistic solution to handwritten digit recognition. This method is promising for various uses in digit recognition scenarios, highlighting its flexibility and usefulness in real-world situations.

Real-Time Navigation Roads: Lightweight and Efficient Convolutional Neural Network (LE-CNN) for Arabic Traffic Sign Recognition in Intelligent Transportation Systems (ITS)

Smart cities are now embracing the new frontier of urban living, with advanced technology being used to enhance the quality of life for residents. Many of these cities have developed transportation systems that improve efficiency and sustainability, as well as quality. Integrating cutting-edge transportation technology and data-driven solutions improves safety, reduces environmental impact, optimizes traffic flow during peak hours, and reduces congestion. Intelligent transportation systems consist of many systems, one of which is traffic sign detection. This type of system utilizes many advanced techniques and technologies, such as machine learning and computer vision techniques. A variety of traffic signs, such as yield signs, stop signs, speed limits, and pedestrian crossings, are among those that the traffic sign detection system is trained to recognize and interpret. Ensuring accurate and robust traffic sign recognition is paramount for the safe deployment of self-driving cars in diverse and challenging environments like the Arab world. However, existing methods often face many challenges, such as variability in the appearance of signs, real-time processing, occlusions that can block signs, low-quality images, and others. This paper introduces an advanced Lightweight and Efficient Convolutional Neural Network (LE-CNN) architecture specifically designed for accurate and real-time Arabic traffic sign classification. The proposed LE-CNN architecture leverages the efficacy of depth-wise separable convolutions and channel pruning to achieve significant performance improvements in both speed and accuracy compared to existing models. An extensive evaluation of the LE-CNN on the Arabic traffic sign dataset that was carried out demonstrates an impressive accuracy of 96.5% while maintaining superior performance with a remarkably low inference time of 1.65 s, crucial for real-time applications in self-driving cars. It achieves high accuracy with low false positive and false negative rates, demonstrating its potential for real-world applications like autonomous driving and advanced driver-assistance systems.

An efficient convolutional neural network for detecting the crime of stealing in videos

In the current age of technological progress, surveillance cameras are extensively placed in densely populated areas to encourage public safety. However, detecting stealing activities within surveillance footage remains a challenge. The global issue of stealing leads to significant casualties and financial losses annually. There has been limited research on efficient techniques to identify this anomaly. In this manuscript, an efficient multi-scale attention-based convolutional neural network with skill optimization algorithm (MSAC-SOA) is proposed for detecting theft during day and night in surveillance videos. This approach integrates with the enhanced single-shot multibook detector (ESSD) for object detection in video footage, facilitating the precise identification and localization of objects within each frame. Subsequently, the proposed MSAC-SOA effectively detects and classifies instances of stealing. Finally, a theft detection alert message is sent to the control room. This approach is assessed with two datasets: University of Central Florida (UCF) crime dataset for daytime theft detection and real-time video data for nighttime theft detection. The proposed method demonstrates faster computation times compared to previous techniques, highlighting its efficiency and suitability in terms of various merits for real-life videos of diverse scenarios. It exhibits robust performance across various metrics, including 99.88% accuracy, 99.66% precision, 100% recall, 99.89% F1-score, 99.83% area under the ROC curve (AUC), and 99.83% specificity.

Classifying Melanoma in ISIC Dermoscopic Images Using Efficient Convolutional Neural Networks and Deep Transfer Learning

Classifying Melanoma in ISIC Dermoscopic Images Using Efficient Convolutional Neural Networks and Deep Transfer Learning

Efficient Bayesian CNN Model Compression using Bayes by Backprop and L1-Norm Regularization

The swift advancement of convolutional neural networks (CNNs) in numerous real-world utilizations urges an elevation in computational cost along with the size of the model. In this context, many researchers steered their focus to eradicate these specific issues by compressing the original CNN models by pruning weights and filters, respectively. As filter pruning has an upper hand over the weight pruning method because filter pruning methods don’t impact sparse connectivity patterns. In this work, we suggested a Bayesian Convolutional Neural Network (BayesCNN) with Variational Inference, which prefaces probability distribution over weights. For the pruning task of Bayesian CNN, we utilized a combined version of L1-norm with capped L1-norm to help epitomize the amount of information that can be extracted through filter and control regularization. In this formation, we pruned unimportant filters directly without any test accuracy loss and achieved a slimmer model with comparative accuracy. The whole process of pruning is iterative and to validate the performance of our proposed work, we utilized several different CNN architectures on the standard classification dataset available. We have compared our results with non-Bayesian CNN models particularly, datasets such as CIFAR-10 on VGG-16, and pruned 75.8% parameters with float-point-operations (FLOPs) reduction of 51.3% without loss of accuracy and has achieved advancement in state-of-art.

Exploring the power of photoplethysmogram matrix for atrial fibrillation detection with integrated explainability

Atrial Fibrillation (AF) detection is paramount for cardiovascular health due to its potential complications. In this study, we investigate the utility of Photoplethysmogram (PPG) for continuous heart rate monitoring and detection of patients with AF. Our approach centres on creating PhotoplethysmoMatrices (PPMs) and leverages Explainable Artificial Intelligence (XAI) techniques to enable accurate AF patient classification with interpretability.Our method involves transforming PPG data into multiple PPM images, which represent aligned peaks within the PPG signal, presented as heatmaps. The diagnostic architecture is a lightweight and efficient Convolutional Neural Network (CNN) combined with attention mechanisms for model transparency.Remarkably, our approach achieves a 100% classification accuracy across multiple experiments, even with variations in the number of users and training images. Furthermore, the attention module underscores the significance of peak positioning and shifting in AF patient detection.Overall, our research makes a substantial contribution to the field of AF patient classification using PPG signals. The combination of image-based preprocessing techniques and explainable architectures enhances accuracy and interpretability, promising improved diagnostic capabilities in clinical settings.

Dataflow optimization with layer-wise design variables estimation method for enflame CNN accelerators

As convolution layers have been proved to be the most time-consuming operation in convolutional neural network (CNN) algorithms, many efficient CNN accelerators have been designed to boost the performance of convolution operations. Previous works on CNN acceleration usually use fixed design variables for diverse convolutional layers, which would lead to inefficient data movements and low utilization of computing resource. We tackle this issue by proposing a flexible dataflow optimization method with design variables estimation for different layers. The optimization method first narrows the design space by the priori constraints, and then enumerates all legal solutions to select the optimal design variables. We demonstrate the effectiveness of the proposed optimization method by implementing representative CNN models (VGG-16, ResNet-18 and MobileNet V1) on Enflame Technology's programmable CNN accelerator, General Computing Unit (GCU). The results indicate that our optimization can significantly enhance the throughput of the convolution layers in ResNet, VGG and MobileNet on GCU, with improvement of up to 1.84×. Furthermore, it achieves up to 2.08× of GCU utilization specifically for the convolution layers of ResNet on GCU.

Best performance with fewest resources: Unveiling the most resource-efficient Convolutional Neural Network for P300 detection with the aid of Explainable AI

Convolutional Neural Networks (CNNs) have shown remarkable prowess in detecting P300, an Event-Related Potential (ERP) crucial in Brain–Computer Interfaces (BCIs). Researchers persistently seek simple and efficient CNNs for P300 detection, exemplified by models like DeepConvNet, EEGNet, and SepConv1D. Noteworthy progress has been made, manifesting in reducing parameters from millions to hundreds while sustaining state-of-the-art performance. However, achieving further simplification or performance improvement beyond SepConv1D appears challenging due to inherent oversimplification. This study explores landmark CNNs and P300 data with the aid of Explainable AI, proposing a simpler yet superior-performing CNN architecture which incorporates (1) precise separable convolution for feature extraction of P300 data, (2) adaptive activation function tailored for P300 data, and (3) customized large learning rate schedules for training P300 data. Termed the Minimalist CNN for P300 detection (P300MCNN), this novel model is characterized by its requirement of the fewest filters and epochs to date, concurrently achieving best performance in cross-subject P300 detection. P300MCNN not only introduces groundbreaking concepts for CNN architectures in P300 detection but also showcases the importance of Explainable AI in demystifying the “black box” design of CNNs.

Detection of Covid-19 based on convolutional neural networks using pre-processed chest X-ray images

The global catastrophe known as COVID-19 has shattered the world’s socioeconomic structure. Effective and affordable diagnosis techniques are crucial for better COVID-19 therapy and the eradication of bogus cases. Due to the daily upsurge in cases, hospitals only have a small supply of COVID-19 test kits. The study describes a deep Convolutional Neural Network (CNN) design for categorizing chest x-ray images in the diagnosis of COVID-19. The lack of a substantial, high-quality chest x-ray picture collection made efficient and exact CNN categorization problematic. The dataset has been pre-processed using an image enhancement strategy to provide an effective training dataset for the proposed CNN model to achieve performance. The proposed model achieves 99.73% of accuracy, 98.95% of specificity, 99.47% of precision, 99.62% of sensitivity, and 98.71% of F1 score. A comparative study between the proposed model and numerous CNN-based COVID-19 detection algorithms is carried out to demonstrate that it outperforms other models. When evaluated on a separate dataset, the suggested model excelled over all other models, generally and explicitly.

Toward Efficient Deep Learning for Graph Drawing (DL4GD).

Due to their great performance in many challenges, Deep Learning (DL) techniques keep gaining popularity in many fields. They have been adapted to process graph data structures to solve various complicated tasks such as graph classification and edge prediction. Eventually, they reached the Graph Drawing (GD) task. This article is an extended version of the previously published (DNN)2 and presents a framework to leverage DL techniques for graph drawing (DL4GD). We demonstrate how it is possible to train a Deep Learning model to extract features from a graph and project them into a graph layout. The method proposes to leverage efficient Convolutional Neural Networks, adapting them to graphs using Graph Convolutions. The graph layout projection is learned by optimizing a cost function that does not require any ground truth layout, as opposed to prior work. This paper also proposes an implementation and benchmark of the framework to study its sensitivity to certain deep learning-related conditions. As the field is novel, and many questions remain to be answered, we do not focus on finding the most optimal implementation of the method, but rather contribute toward a better understanding of the approach potential. More precisely, we study different learning strategies relative to the models training datasets. Finally, we discuss the main advantages and limitations of DL4GD.

A deep learning based architecture for malaria parasite detection

During last decade, medical imaging has attracted great deal of research interests. Deep learning applications has revolutionized medical image analysis and diseases diagnosis. Convolutional neural networks (CNNs)-a class of deep learning-have been widely used for classification and feature extraction, and they revealed good performance for various imaging applications. However, despite the advances in medicine, malaria remains among the world’s deadliest diseases. Only in 2020, malaria recorded 241 million clinical episodes, and 627,000 deaths. The disease is examined visually through a microscope, which depends on the pathologists experience and skills and results may vary in different laboratories. This paper proposes an efficient CNN architecture that could be used in diagnosing of malaria disease. By processing on 27,558 red blood smear cell images with balanced samples of parasitized and unparasitized cells on a publicly available malaria dataset from the National Institute of Health, the proposed model achieves high accuracy rate with 99.8%, 98.2, and 97.7% for training, validation and testing sets. Furthermore, the statistical results approve that the proposed model is outperforming the state-of-the-art models.

Data augmentation for 3D seismic fault interpretation using deep learning

Manual seismic interpretation of faults is a tedious and complicated process, which is prone to human error and bias. A semi-automatic approach for interpreting faults on seismic data is to use attributes that highlight discontinuities. These methods need to be optimized by the interpreter, thus constantly rely on the interpreter's knowledge. Recently, Machine Learning (ML) techniques in general and Convolutional Neural Networks (CNN) as part of Deep Neural Networks (DNN), in particular, have been used to detect and image faults on seismic data in order to make the process more automated and accurate. Supervised CNN learns and evolves from manually annotated or labeled fault interpretations.In this study, we have applied supervised 2D CNN to image faults on seismic data through binary segmentation. The study was performed on 3D marine seismic surveys in off-shore Norway collected from three separate locations along the Norwegian Continental Shelf, utilizing Efficient UNET CNN architecture. We applied data augmentation (geometric transformations) and hyperparameter tuning to improve the learning process and performance of the deep learning algorithm.We used noise content, acquisition type, imaging type, and fault scale as the main criteria to select seismic surveys. The application of data augmentation to the training and testing data generally led to improvement in the performance of 2D CNN on fault predictions, although the amount of improvement varied with respect to different surveys. The initial 2D CNN fault prediction improvement mainly relied on the quality, and size of faults present in the 3D seismic volumes. Further, improvement was achieved by the adjustment of certain hyperparameters affecting the training and testing process of the 2D CNN. However, we found little to no improvement on one seismic volume containing high levels of noise.

A Visual Computing Unified Application Using Deep Learning and Computer Vision Techniques

Vision Studio aims to utilize a diverse range of modern deep learning and computer vision principles and techniques to provide a broad array of functionalities in image and video processing. Deep learning is a distinct class of machine learning algorithms that utilize multiple layers to gradually extract more advanced features from raw input. This is beneficial when using a matrix as input for pixels in a photo or frames in a video. Computer vision is a field of artificial intelligence that teaches computers to interpret and comprehend the visual domain. The main functions implemented include deepfake creation, digital ageing (de-ageing), image animation, and deepfake detection. Deepfake creation allows users to utilize deep learning methods, particularly autoencoders, to overlay source images onto a target video. This creates a video of the source person imitating or saying things that the target person does. Digital aging utilizes generative adversarial networks (GANs) to digitally simulate the aging process of an individual. Image animation utilizes first-order motion models to create highly realistic animations from a source image and driving video. Deepfake detection is achieved by using advanced and highly efficient convolutional neural networks (CNNs), primarily employing the EfficientNet family of models.

Research on Walnut (Juglans regia L.) Classification Based on Convolutional Neural Networks and Landsat-8 Remote Sensing Imagery

The study explores the use of convolutional neural networks (CNNs) and satellite remote sensing imagery for walnut analysis in Ganquan Township, Alar City, Xinjiang. The recent growth of walnut cultivation in Xinjiang presents challenges for manual data collection, making satellite imagery and computer vision algorithms a practical solution. Landsat-8 satellite images from Google Earth Engine underwent preprocessing, and experiments were conducted to enhance the ResNet model, resulting in improved accuracy and efficiency. Experiments were conducted to evaluate multiple CNN models and traditional methods, and the best detection method was chosen through comparisons. A comparison was drawn between traditional algorithms and convolutional neural network algorithms based on metrics such as precision, recall, f1-score, accuracy, and total time. The results indicated that although traditional methods were more efficient compared to CNN, they exhibited lower accuracy. In the context of this research, prioritizing efficiency at the cost of accuracy was deemed undesirable. Among the traditional algorithms employed in this study, k-NN produced the most favorable outcomes, with precision, recall, f1-score, and accuracy reaching 75.78%, 92.43%, 83.28%, and 84.46%, respectively, although these values were relatively lower than those of the CNN algorithm models. Within the CNN models, the ResNet model demonstrated superior performance, yielding corresponding results of 92.47%, 94.29%, 93.37%, and 93.27%. The EfficientNetV2 model also displayed commendable results, with precision, recall, and f1-score achieving 96.35%, 91.44%, and 93.83%. Nevertheless, it is worth noting that the classification efficiency of EfficientNetV2 fell significantly short of that of ResNet. Consequently, in this study, the ResNet model proved to be relatively more effective. Once optimized, the most efficient CNN model closely rivals traditional algorithms in terms of time efficiency for generating results while significantly surpassing them in accuracy. Through our studies, we discovered that once optimized, the most efficient CNN model closely rivals traditional algorithms in terms of time efficiency for generating results while significantly surpassing them in accuracy. In this study, empirical evidence demonstrates that integrating CNN-based methods with satellite remote sensing technology can effectively enhance the statistical efficiency of agriculture and forestry sectors, thus leading to substantial reductions in operational costs. These findings lay a solid foundation for further research in this field and offer valuable insights for other agricultural and forestry-related studies.

Toward Efficient Convolutional Neural Networks With Structured Ternary Patterns.

High-efficiency deep learning (DL) models are necessary not only to facilitate their use in devices with limited resources but also to improve resources required for training. Convolutional neural networks (ConvNets) typically exert severe demands on local device resources and this conventionally limits their adoption within mobile and embedded platforms. This brief presents work toward utilizing static convolutional filters generated from the space of local binary patterns (LBPs) and Haar features to design efficient ConvNet architectures. These are referred to as Structured Ternary Patterns (STePs) and can be generated during network initialization in a systematic way instead of having learnable weight parameters thus reducing the total weight updates. The ternary values require significantly less storage and with the appropriate low-level implementation, can also lead to inference improvements. The proposed approach is validated using four image classification datasets, demonstrating that common network backbones can be made more efficient and provide competitive results. It is also demonstrated that it is possible to generate completely custom STeP-based networks that provide good trade-offs for on-device applications such as unmanned aerial vehicle (UAV)-based aerial vehicle detection. The experimental results show that the proposed method maintains high detection accuracy while reducing the trainable parameters by 40%-80%. This work motivates further research toward good priors for nonlearnable weights that can make DL architectures more efficient without having to alter the network during or after training.

Identification of CNN hyper-parameters for tobacco leaf quality classification on Nvidia Jetson Nano

At the moment, there are some inaccuracies in manual classification for tobacco leaf quality selection, which are influenced by some factors, such as human fatigue or poor lighting. These cases lead to the need for another method that is more consistent, faster and reliable. This research is an implementation of CNN (Convolutional Neural Network) in the classification of fresh tobacco leaves in terms of maturity grades. The primary objective is to develop an efficient CNN model capable to automize the classification of tobacco leaves into three maturity criteria: immature, mature, and old. This methodology consists of some key factors, including color thresholding strategies to purge the noise from the background, Basic Image Manipulation approaches, the systemized screening of different input sizes, and CNN models to enhance the results. The result of this research proves that the developed CNN model has 97.9 % accuracy achieved following 200 training sessions. The model is trained on a dataset comprising 1,249 fresh leaf photos, with a balanced 80:10:10 for train, validation and test ratio. However, the study emphasizes that the CNN model has successfully supported the tobacco leaf discrimination on a Jetson Nano Single-Board Computer with a Graphic Processing Unit (GPU). The study extends beyond the mere theoretical contribution to practical applications in sorting “Gagang Rejeb Sidi” tobacco leaf, the highest quality tobacco variety in South Malang, East Java, Indonesia. Classification using a webcam as an input device shows the fastest processing time of 203.17 ms and the maximum is 1,363 ms. This CNN model algorithm will be applied to a tobacco leaf selector machine, which has a high-speed conveyor and a three-position selector arm. The machine will be operated close to the field in post-harvest time under uniform lighting conditions. Overall, the result of this research is highly relevant in terms of the short duration and accuracy for understanding the commodity classification. It provides a new angle toward speeding up the classification process and improving Indonesian tobacco quality

Bare‐Bones particle Swarm optimization‐based quantization for fast and energy efficient convolutional neural networks

AbstractNeural network quantization is a critical method for reducing memory usage and computational complexity in deep learning models, making them more suitable for deployment on resource‐constrained devices. In this article, we propose a method called BBPSO‐Quantizer, which utilizes an enhanced Bare‐Bones Particle Swarm Optimization algorithm, to address the challenging problem of mixed precision quantization of convolutional neural networks (CNNs). Our proposed algorithm leverages a new population initialization, a robust screening process, and a local search strategy to improve the search performance and guide the population towards a feasible region. Additionally, Deb's constraint handling method is incorporated to ensure that the optimized solutions satisfy the functional constraints. The effectiveness of our BBPSO‐Quantizer is evaluated on various state‐of‐the‐art CNN architectures, including VGG, DenseNet, ResNet, and MobileNetV2, using CIFAR‐10, CIFAR‐100, and Tiny ImageNet datasets. Comparative results demonstrate that our method delivers an excellent tradeoff between accuracy and computational efficiency.

Analysis for diagnosis of pneumonia symptoms using chest X-ray based on MobileNetV2 models with image enhancement using white balance and contrast limited adaptive histogram equalization (CLAHE)

This study focuses on diagnosing pneumonia symptoms using chest X-ray (CXR) images. It employs the MobileNetV2 model alongside image enhancement techniques, including white balance and CLAHE. MobileNetV2 is a computationally efficient Convolutional Neural Network (CNN) known for its robust image recognition capabilities. White balance corrects color imbalances in CXR images, ensuring color consistency, while CLAHE enhances contrast and image details for improved analysis. The enhanced CXR images are fed into the pre-trained MobileNetV2 model, which distinguishes pneumonia and non-pneumonia cases. The study aims to enhance pneumonia diagnosis accuracy, benefiting from MobileNetV2’s efficiency and image enhancement. Notably, it achieved high accuracy and low loss for both three-class (91.17% accuracy, 35.0% loss) and two-class (99.76% accuracy, 7% loss) classifications, with the best results in the 50-epoch test. However, it is essential to consider the trade-off between computing time and the risk of overfitting when increasing epochs. Future research could explore additional features to further enhance model performance.