Lightweight Convolutional Neural Network Model Research Articles

A novel imbalanced dataset mitigation method and ECG classification model based on combined 1D_CBAM-autoencoder and lightweight CNN model

To address the problems of the imbalanced datasets, limited computing power, and memory capability of wearable devices, this paper proposes a new hybrid method for imbalanced dataset mitigation and a novel lightweight Convolutional Neural Network (CNN) framework for electrocardiogram(ECG) signal classification. First, data enhancement method is used by changing the amplitude of a small number of samples and combining with the weighted loss function in the deep learning model. Then, One-Dimensional Convolutional Block Attention Module (1D_CBAM) mechanism autoencoder model is proposed to compress the ECG heartbeat signals and extract the features of the ECG heartbeats. A modified one-dimensional lightweight CNN model based on MobileNet architecture is eventually used to categorize various arrhythmias utilizing the above features. Using the MIT-BIH arrhythmia dataset, the proposed model is tested by dividing the MIT Database(MITDB) into Dataset1 and Dataset2, in conjunction with suggestions from AAMI and other heartbeat signal classification methods. The experimental results show that the proposed imbalanced dataset mitigation method has higher performance than other methods in this paper, the proposed model has the most stable convergence and the lowest reconstruction loss (approximately 0.1) in the feature extraction process under the same conditions. During the classification process, the lightweight CNN model has approximately 4x fewer trainable parameters(Params: 12,788) and its performance [Dataset1: F1-Score: 96.24% ACC: 98.21%; Dataset2: F1-Score: 98.73% ACC: 99.61%] is comparable to the traditional CNN model [Dataset1: F1-Score: 96.98% ACC: 99.17%; Dataset2: F1-Score: 99.18% ACC: 99.72%].

An extremely lightweight CNN model for the diagnosis of chest radiographs in resource-constrained environments.

In recent years, deep learning methods have been successfully used for chest x-ray diagnosis. However, such deep learning models often contain millions of trainable parameters and have high computation demands. As a result, providing the benefits of cutting-edge deep learning technology to areas with low computational resources would not be easy. Computationally lightweight deep learning models may potentially alleviate this problem. We aim to create a computationally lightweight model for the diagnosis of chest radiographs. Our model has only 0.14M parameters and 550 KB size. These make the proposed model potentially useful for deployment in resource-constrainedenvironments. We fuse the concept of depthwise convolutions with squeeze and expand blocks to design the proposed architecture. The basic building block of our model is called Depthwise Convolution In Squeeze and Expand (DCISE) block. Using these DCISE blocks, we design an extremely lightweight convolutional neural network model (ExLNet), a computationally lightweight convolutional neural network (CNN) model for chest x-raydiagnosis. We perform rigorous experiments on three publicly available datasets, namely, National Institutes of Health (NIH), VinBig ,and Chexpert for binary and multi-class classification tasks. We train the proposed architecture on NIH dataset and evaluate the performance on VinBig and Chexpert datasets. The proposed method outperforms several state-of-the-art approaches for both binary and multi-class classification tasks despite having a significantly less number ofparameters. We design a lightweight CNN architecture for the chest x-ray classification task by introducing ExLNet which uses a novel DCISE blocks to reduce the computational burden. We show the effectiveness of the proposed architecture through various experiments performed on publicly available datasets. The proposed architecture shows consistent performance in binary as well as multi-class classification tasks and outperforms other lightweight CNN architectures. Due to a significant reduction in the computational requirements, our method can be useful for resource-constrained clinical environment aswell.

Drivable Area Detection in Unstructured Environments based on Lightweight Convolutional Neural Network for Autonomous Driving Car

Road detection technology is an important part of the automatic driving environment perception system. With the development of technology, the situations that automatic driving needs to consider will become broader and more complex. This paper contributes a lightweight convolutional neural network model, incorporating novel convolution and parallel pooling modules, an improved network activation function, and comprehensive training and verification with multiple datasets. The proposed model achieves high accuracy in detecting drivable areas in complex autonomous driving situations while significantly improving real-time performance. In addition, we collect data in the field and create small datasets as reference datasets for testing algorithms. This paper designs relevant experimental scenarios based on the datasets and experimental platforms and conducts simulations and real-world vehicle experiments to verify the effectiveness and stability of the algorithm models and technical solutions. The method achieves an MIoU of 90.19 and a single batch time of 340 ms with a batch size of 8, which substantially reduces the runtime relative to a typical deep network structure like ResNet50.

English handwriting recognition based on the convolutional neural network

Handwriting recognition is widely used in the retrieval, recognition, and management of various information to improve efficiency in various industries. Convolutional neural networks help people get rid of the feature of extracting information feature sets manually and significantly improve recognition efficiency. In this paper, a generalization-enhanced network recognition model is proposed using an improved lightweight convolutional neural network model. An improved method is used to adapt word recognition by the idea of pre-recognition segmentation. In addition, the generalization is enhanced by diversifying and pre-processing the dataset so that the algorithm can obtain noise-resistant performance and detail retention and allow the recognition system to recognize various types of scenes. The results show that the model achieves an accuracy of 93% on the test set. Compared with other classical network models, the model has higher recognition accuracy, faster convergence, and better generalization ability. The system elaborated in this paper can be used for devices with weak computer processing power.

Vision-based underwater target real-time detection for autonomous underwater vehicle subsea exploration

Autonomous underwater vehicles (AUVs) equipped with online visual inspection systems can detect underwater targets during underwater operations, which is of great significance to subsea exploration. However, the undersea scene has some instinctive challenging problems, such as poor lighting conditions, sediment burial, and marine biofouling mimicry, which makes it difficult for traditional target detection algorithms to achieve online, reliable, and accurate detection of underwater targets. To solve the above issues, this paper proposes a real-time object detection algorithm for underwater targets based on a lightweight convolutional neural network model. To improve the imaging quality of underwater images, contrast limited adaptive histogram equalization with the fused multicolor space (FCLAHE) model is designed to enhance the image quality of underwater targets. Afterwards, a spindle-shaped backbone network is designed. The inverted residual block and group convolutions are used to extract depth features to ensure the target detection accuracy on one hand and to reduce the model parameter volume on the other hand under complex scenarios. Through extensive experiments, the precision, recall, and mAP of the proposed algorithm reached 91.2%, 90.1%, and 88.3%, respectively. It is also noticeable that the proposed method has been integrated into the embedded GPU platform and deployed in the AUV system in the practical scenarios. The average computational time is 0.053s, which satisfies the requirements of real-time object detection.

AI‐enabled farm‐friendly automatic machine for washing, image‐based sorting, and weight grading of citrus fruits: Design optimization, performance evaluation, and ergonomic assessment

AbstractThe modernization of postharvest operations and penetration of emerging technologies in horticultural processing have provided intelligent solutions for reducing postharvest losses. Work environmental and occupational health issues require immediate attention as the awkward posture and continuous drudgery‐prone on‐farm sorting and grading activities may lead to musculoskeletal disorders. The main objective of this study was to develop an automatic farm‐friendly machine for real‐time citrus fruit washing, image‐based sorting, and weight grading; designed optimally and equipped with an embedded system comprising a lightweight convolutional neural network (CNN) model. Also included in this study was a thorough ergonomic assessment of the developed machine in a real work environment. The parametric choice of the fruit washing and singulation system was performed by employing computational fluid dynamics modeling and response surface methodology designed optimization. It was observed that under steady‐state conditions, the water jet would arrive at a velocity of 11.36 m/s which would eventually suit a singulation conveyor with a slope of 25°. A noninvasive grading and sorting approach for citrus fruits is presented in this paper that leverages deep learning to classify the fruits into “accept” and “reject” classes. The custom lightweight CNN model “SortNet” has shown excellent classification results with an overall accuracy of 97.6%. The ergonomic evaluation shows that the average body part discomfort score in case of operating an automatic fruit grading machine was much lower (12.3 ± 2.0) than the traditional method (30.9 ± 3.3). Further, in the case of machine operation, the percentage load on the muscles ranged from 28.67 to 34.31 reflecting that subjects can work for longer duration on the machine without fatigue as compared with the traditional manual operation.

Deep convolutional neural network-based effective model for 2D ear recognition using data augmentation

ABSTRACT In the pandemic of COVID-19, identifying a person from their face became difficult due to wearing of mask. In regard to the given circumstances, the authors have remarkably put effort on identifying a person using 2D ear images based on deep convolutional neural network (CNNs). They investigated the challenges of limited data and varying environmental conditions in this regards. To deal with such challenges, the authors developed an augmentation-based light-weight CNN model using CPU enabled machine so that it can be ported into embedded devices. While applying data augmentation technique to enhance the quality and size of training dataset, the authors analysed and discussed the different augmentation parameters (rotation, flipping, zooming, and fill mode) that are effective for generating the large number of sample images of different variability. The model works well on constrained and unconstrained ear datasets and achieves good recognition accuracy. It also reduces the problem of overfitting.

Patient-independent seizure detection based on long-term iEEG and a novel lightweight CNN

Objective. Patient-dependent seizure detection based on intracranial electroencephalography (iEEG) has made significant progress. However, due to the difference in the locations and number of iEEG electrodes used for each patient, patient-independent seizure detection based on iEEG has not been carried out. Additionally, current seizure detection algorithms based on deep learning have outperformed traditional machine learning algorithms in many performance metrics. However, they still have shortcomings of large memory footprints and slow inference speed. Approach. To solve the above problems of the current study, we propose a novel lightweight convolutional neural network model combining the Convolutional Block Attention Module (CBAM). Its performance for patient-independent seizure detection is evaluated on two long-term continuous iEEG datasets: SWEC-ETHZ and TJU-HH. Finally, we reproduce four other patient-independent methods to compare with our method and calculate the memory footprints and inference speed for all methods. Main results. Our method achieves 83.81% sensitivity (SEN) and 85.4% specificity (SPE) on the SWEC-ETHZ dataset and 86.63% SEN and 92.21% SPE on the TJU-HH dataset. In particular, it takes only 11 ms to infer 10 min iEEG (128 channels), and its memory footprint is only 22 kB. Compared to baseline methods, our method not only achieves better patient-independent seizure detection performance but also has a smaller memory footprint and faster inference speed. Significance. To our knowledge, this is the first iEEG-based patient-independent seizure detection study. This facilitates the application of seizure detection algorithms to the future clinic.

A Lightweight Neural Network-Based Method for Identifying Early-Blight and Late-Blight Leaves of Potato

Crop pests and diseases are one of the most critical disasters that limit agricultural production. In this paper, we trained a lightweight convolutional neural network model and built a Django framework-based potato disease leaf recognition system, which can recognize three types of potato leaf images including early blight, late blight, and healthy. A lightweight, neural network-based model for the identification of early potato leaf diseases significantly reduces the number of model parameters, whereas the accuracy of Top-1 identification is over 93%. We imported the trained model into the Django framework to build a website for a potato early leaf disease identification system, thus providing technical support for the implementation of a mobile-based potato leaf disease identification and early warning system.

Optimized pointwise convolution operation by Ghost blocks

<abstract> <p>In the lightweight convolutional neural network model, the pointwise convolutional structure occupies most of the parameters and computation amount of the model. Therefore, improving the pointwise convolution structure is the best choice to optimize the lightweight model. Aiming at the problem that the pointwise convolution in MobileNetV1 and MobileNetV2 consumes too many computation resources, we designed the novel Ghost-PE and Ghost-PC blocks. First, in order to optimize the channel expanded pointwise convolution with the number of input channels less than the output, Ghost-PE makes full use of the feature maps generated by main convolution of the Ghost module, and adds global average pooling and depth convolution operation to enhance the information of feature maps generated through cheap convolution. Second, in order to optimize the channel compressed pointwise convolution with the number of input channels more than the output, Ghost-PC adjusts the Ghost-PE block to make full use of the features generated by cheap convolution to enhance the feature channel information. Finally, we optimized MobileNetV1 and MobileNetV2 models by Ghost-PC and Ghost-PE blocks, and then tested on Food-101, CIFAR and Mini-ImageNet datasets. Compared with other methods, the experimental results show that Ghost-PE and Ghost-PC still maintain a relatively high accuracy in the case of a small number of parameters.</p> </abstract>

Light weight convolutional neural network and low-dimensional images transformation approach for classification of thermal images

Thermal energy is emitted in the infrared range between X-ray and Gamma rays, which are invisible to the human eye. Thermal cameras can detect the temperature that arises due to the heat emitted by the objects in a non-contact way and transform it into an image. These images ensure to detection of objects regardless of ambient occlusion. Based on this problem, five different classification models were proposed within the scope of the study. New low-dimensional images were obtained by extracting the features of thermal images with HOG (Histogram Oriented of Gradients), LBP (Local Binary Pattern), SIFT (Scale Invariant Feature Transform), and GF (Gabor Filter) methods. These images are classified by a CNN (Convolutional Neural Network) model called LW-CNN (Light Weight CNN). Raw thermal images were classified with the LW-CNN model without pre-processing. In order to analyze the efficiency of the proposed models, the results were compared via the pre-trained VGG16 model. Three different datasets containing thermal images were used in classification processes. The highest classification accuracy was obtained from the LW-CNN model in the performance evaluations carried out on the three datasets. With this model, the classification accuracies obtained from the datasets are 98.58%, 95.56%, and 100%, respectively.

DFCANet: A Novel Lightweight Convolutional Neural Network Model for Corn Disease Identification

The identification of corn leaf diseases in a real field environment faces several difficulties, such as complex background disturbances, variations and irregularities in the lesion areas, and large intra-class and small inter-class disparities. Traditional Convolutional Neural Network (CNN) models have a low recognition accuracy and a large number of parameters. In this study, a lightweight corn disease identification model called DFCANet (Double Fusion block with Coordinate Attention Network) is proposed. The DFCANet consists mainly of two components: The dual feature fusion with coordinate attention and the Down-Sampling (DS) modules. The DFCA block contains dual feature fusion and Coordinate Attention (CA) modules. In order to completely fuse the shallow and deep features, these features were fused twice. The CA module suppresses the background noise and focuses on the diseased area. In addition, the DS module is used for down-sampling. It reduces the loss of information by expanding the feature channel dimension and the Depthwise convolution. The results show that DFCANet has an average recognition accuracy of 98.47%. It is more efficient at identifying corn leaf diseases in real scene images, compared with VGG16 (96.63%), ResNet50 (93.27%), EffcientNet-B0 (97.24%), ConvNeXt-B (94.18%), DenseNet121 (95.71%), MobileNet-V2 (95.41%), MobileNetv3-Large (96.33%), and ShuffleNetV2-1.0× (94.80%) methods. Moreover, the model’s Params and Flops are 1.91M and 309.1M, respectively, which are lower than heavyweight network models and most lightweight network models. In general, this study provides a novel, lightweight, and efficient convolutional neural network model for corn disease identification.

EdgeFireSmoke: A Novel Lightweight CNN Model for Real-Time Video Fire–Smoke Detection

The planet Earth is being affected by a series of wildfires, which have been steadily increasing over the last two decades. Forests have undergone deforestation due to natural forest fires and forest fires caused by man. These events are occurring on a global scale, and in Brazil, these wildfires are having an extreme impact on the Amazon forest as well as other forest biomes. This article proposes a novel lightweight convolutional neural network (CNN) model for wildfire detection through RGB images. This new method presents more advantages than the other methods used for the same task. Our CNN architecture can be used with aerial images from unmanned aerial vehicles and from video surveillance systems, combined with edge computing devices for image processing with a CNN. The proposed system is able to send wildfire alerts. The images do not have to be sent to a cloud computer as they can be processed in an edge device. However, it sends a string of alerts whenever a wildfire is detected. The evaluation of our proposed method showed that it required about 30 ms for the classification time, per image, and achieved an accuracy of 98.97% and an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F1$</tex-math></inline-formula> -score of 95.77%, which is a very promising result.

Classification of UAVs Utilizing Fixed Boundary Empirical Wavelet Sub-Bands of RF Fingerprints and Deep Convolutional Neural Network

Unmanned aerial vehicle (UAV) classification and identification have many applications in a variety of fields, including UAV tracking systems, antidrone systems, intrusion detection systems, military, space research, product delivery, agriculture, search and rescue, and internet carrier. It is challenging to identify a specific drone and/or type in critical scenarios, such as intrusion. In this article, a UAV classification method that utilizes fixed boundary empirical wavelet sub-bands of radio frequency (RF) fingerprints and a deep convolutional neural network (CNN) is proposed. In the proposed method, RF fingerprints collected from UAV receivers are decomposed into 16 fixed boundary empirical wavelet sub-band signals. Then, these sub-band signals are then fed into a lightweight deep CNN model to classify various types of UAVs. Using the proposed method, we classify a total of 15 different commercially available UAVs with an average testing accuracy of 97.25%. The proposed model is also tested with various sampling points in the signal. Furthermore, the proposed method is compared with recently reported works for classifying UAVs utilizing remote controller RF signals.

A lightweight convolutional neural network model with receptive field block for C-shaped root canal detection in mandibular second molars

Rapid and accurate detection of a C-shaped root canal on mandibular second molars can assist dentists in diagnosis and treatment. Oral panoramic radiography is one of the most effective methods of determining the root canal of teeth. There are already some traditional methods based on deep learning to learn the characteristics of C-shaped root canal tooth images. However, previous studies have shown that the accuracy of detecting the C-shaped root canal still needs to be improved. And it is not suitable for implementing these network structures with limited hardware resources. In this paper, a new lightweight convolutional neural network is designed, which combined with receptive field block (RFB) for optimizing feature extraction. In order to optimize the hardware resource requirements of the model, a lightweight, multi-branch, convolutional neural network model was developed in this study. To improve the feature extraction ability of the model for C-shaped root canal tooth images, RFB has been merged with this model. RFB has achieved excellent results in target detection and classification. In the multiscale receptive field block, some small convolution kernels are used to replace the large convolution kernels, which allows the model to extract detailed features and reduce the computational complexity. Finally, the accuracy and area under receiver operating characteristics curve (AUC) values of C-shaped root canals on the image data of our mandibular second molars were 0.9838 and 0.996, respectively. The results show that the deep learning model proposed in this paper is more accurate and has lower computational complexity than many other similar studies. In addition, score-weighted class activation maps (Score-CAM) were generated to localize the internal structure that contributed to the predictions.

Development of Light-Weight Convolutional Neural Network Model to Diagnose Tuberculosis

Tuberculosis (TB) is a contagious chest infection. World Health Organization has introduced different TB control programs in various countries. For the diagnosis of TB, the doctors mostly recommend chest X-ray (CXR) because it is more cost-effective and less time-consuming than existing sputum tests and Tuberculosis Skin Tests (TST). As per the research, deep learning models are best for TB diagnosis, by using CXR rather than normal eye-sight-based traditional method. Since doctor’s eye-sight or his experience is prone to human error, therefore, to solve this problem many Convolutional Neural Network (CNN) based models are introduced. Some of these models have high computational costs, and better accuracy making them heavy model. Whereas, others have less computational costs and lower accuracy making them light-weight models. Such models are further modified by the researchers to be more appropriate for better TB diagnosis, termed as Transfer Learning (TL) technique. However, TL leads to complex CNN structure and high computational cost. The proposed model named as Light TBNET(L-TBNET), attempts to provide less computational costs and higher accuracy simultaneously, as compared to other models such as, ShuffleNet, ResNet-50, MobileNet v2, Inception, and DenseNet. Moreover, the proposed does not include TL technique. This is accomplished by combining standard convolutional layers as well as depth-wise separable convolutional layers resulting in a hybrid model. The accuracy of the proposed model is 96% with lesser computational cost. In this way, the model contributes in providing a light-weight CNN model with higher accuracy.

Ship Fire Detection Based on an Improved YOLO Algorithm with a Lightweight Convolutional Neural Network Model.

Ship fire is one of the greatest dangers to ship navigation safety. Nevertheless, typical detection methods have limited detection effectiveness and accuracy due to distance restrictions and ship motion. Although the issue can be addressed by image recognition algorithms based on deep learning, the computational complexity and efficiency for ship detection are tough. This paper proposes a lightweight target identification technique based on the modified YOLOv4-tiny algorithm for the precise and efficient detection of ship fires, taking into account the distinctive characteristics of ship fires and the marine environment. Initially, a multi-scale detection technique is applied to broaden the detection range and integrate deep semantic information, thereby enhancing the feature information of small targets and obscured objects and improving the detection precision. Then, the proposed algorithm employs the SE attention mechanism for inter-channel feature fusion to improve the capability of feature extraction and the precision of ship fire detection. Last but not least, picture transformation and migration learning are added to the small ship fire dataset to accelerate the convergence pace, improve the convergence effect, and reduce dataset dependence. The simulation experiments reveal that the proposed I-YOLOv4-tiny + SE model outperforms the benchmark algorithm in terms of ship fire detection accuracy and detection efficiency and that it satisfies the real-time ship fire warning criteria in demanding maritime environments.

IT Monitoring and Management of Learning Quality of Online Courses for College Students Based on Machine Learning

Network information technology and distance technology learning provide convenience for college students to learn online courses, but some problems have also been found in practice, and schools need to pay attention to improving students’ learning quality and supervision. The cross-spatial nature of the study can be used to study how to detect students’ learning fatigue and learning concentration in online classrooms. This paper first designs a lightweight convolutional neural network model for eye state classification and verifies the performance of the model. The designed model has a compact structure and a high recognition rate. Combined with the human eye positioning algorithm, the recognition of the opening and closing state of the eyes is realized. Finally, the feasibility of using the PERCLOS value for fatigue detection, Euler pitch angle, and yaw is verified by experiments. Corners can be used to detect student attention. The method can enhance the synergistic supervision of other cooperative methods, thus improving the quality and effectiveness of online learning for college students, promoting the development of digital modern teaching and learning management models, and exploring possible future technologies and corresponding changes in teaching methods and management models.

CodnNet: A lightweight CNN architecture for detection of COVID-19 infection

The application of Convolutional Neural Network (CNN) on the detection of COVID-19 infection has yielded favorable results. However, with excessive model parameters, the CNN detection of COVID-19 is low in recall, highly complex in computation. In this paper, a novel lightweight CNN model, CodnNet is proposed for quick detection of COVID-19 infection. CodnNet builds a more effective dense connections based on DenseNet network to make features highly reusable and enhances interactivity of local and global features. It also uses depthwise separable convolution with large convolution kernels instead of traditional convolution to improve the range of receptive field and enhances classification performance while reducing model complexity. The 5-Fold cross validation results on Kaggle’s COVID-19 Dataset showed that CodnNet has an average precision of 97.9%, recall of 97.4%, F1score of 97.7%, accuracy of 98.5%, mAP of 99.3%, and mAUC of 99.7%. Compared to the typical CNNs, CodnNet with fewer parameters and lower computational complexity has achieved better classification accuracy and generalization performance. Therefore, the CodnNet model provides a good reference for quick detection of COVID-19 infection.

AI-Assisted Edge Vision for Violence Detection in IoT-Based Industrial Surveillance Networks

Analyzing surveillance videos is mandatory for the public and industrial security. Overwhelming growth in computer vision fields has been made to automate the surveillance system in terms of human activity recognition, such as behavior analysis and violence detection (VD). However, it is challenging to detect and analyze the violent scenes intelligently to fulfill the notion of Industrial Internet of Things (IIoT)-based surveillance buoyed by constrained resources to reduce computational power. To tackle this challenge, in this article, an artificial intelligence enabled IIoT-based framework with VD-Network (VD-Net) is proposed. First, the input video frames are passed to light-weight convolutional neural network model for important information collection including humans or suspicious objects such as knives/guns. Upon suspicious object detection, an alert is generated as an earlier VD in IIoT network while the information is shared with concern departments. Only the frames with objects are forwarded to cloud for detail investigation where features are extracted using convolutional long short-term memory (ConvLSTM). The latter from ConvLSTM is propagated to gated recurrent unit for final VD. The conducted experiments and ablation study on the existing surveillance and nonsurveillance datasets empirically validate the effectiveness of the proposed VD-Net by improving 3.9% increase in the accuracy compared with the state-of-the-art VD methods.