VGG16 Network Research Articles

Improved Faster R-CNN-Based Anomaly Target Detection for Truck Driving Environment

With the development of the economy and the improvement of people's living standards, the volume of goods transported is constantly on the rise. At present, the dominant position in the transport industry is still road transport, and at the same time the problem of road traffic accidents is becoming increasingly prominent. Generally speaking, the truck is relatively large, with a wide blind spot and high cab height, which makes the driver unable to have a good observation and control of the environment around the vehicle, so it is extremely easy to have all kinds of traffic accidents. In such a context, an improved Faster R-CNN algorithm is proposed for the detection of abnormal targets in the truck driving environment. First, the ResNet-50 network is chosen to replace the VGG network, which reduces the training difficulty and effectively improves the gradient disappearance problem; then, in order to enhance the feature extraction ability of the network, the Squeeze-and-Excitation attention mechanism is introduced in the residual structure to strengthen the feature extraction ability; finally, the original feature pyramid network structure is improved by adding a self Finally, the original feature pyramid structure was improved by adding a bottom-up channel enhancement route to improve the propagation of lower-level features and further enhance the propagation of feature information. The experimental results show that the improved mAP value improves by 2.05% compared to the original algorithm.

SSD Object Detection Algorithm Based on Feature Fusion and Channel Attention

Abstract Aiming at the problems of low object detection accuracy due to complex background and insufficient semantic information of shallow features in the object detection SSD algorithm, this paper improves the existing SSD algorithm. First, the original vgg16 network is replaced by the ResNet50 network, and the residual network structure as well as the Batch Normalization layer are added, which are used to improve the accuracy of the feature extraction network; Second, a feature fusion module is designed to fuse adjacent feature maps to improve the detection effect by integrating contextual information; Third, the SE attention mechanism is introduced to give channel weights adaptively and enhance the useful feature channels; Finally, the object detection analysis experiments are conducted on the PASCAL VOC2012 dataset. The experimental results show that the improved SSD algorithm in this paper is able to achieve an mean average precision of 72.7% in the data set, which is 2.1% better than the original SSD-VGG16 and greatly improves the object detection effect.

Segmentation of Remote Sensing Images Based on U-Net Multi-Task Learning

In order to accurately segment architectural features in high-resolution remote sensing images, a semantic segmentation method based on U-net network multi-task learning is proposed. First, a boundary distance map was generated based on the remote sensing image of the ground truth map of the building. The remote sensing image and its truth map were used as the input in the U-net network, followed by the addition of the building ground prediction layer at the end of the U-net network. Based on the ResNet network, a multi-task network with the boundary distance prediction layer was built. Experiments involving the ISPRS aerial remote sensing image building and feature annotation data set show that compared with the full convolutional network combined with the multi-layer perceptron method, the intersection ratio of VGG16 network, VGG16 + boundary prediction, ResNet50 and the method in this paper were increased by 5.15%, 6.946%, 6.41% and 7.86%. The accuracy of the networks was increased to 94.71%, 95.39%, 95.30% and 96.10% respectively, which resulted in high-precision extraction of building features.

High-Performance Visual Tracking Based on High-Order Pooling Network

Convolution Neural Network (CNN) features have been widely used in visual tracking due to their powerful representation. As an important component of CNN, the pooling layer plays a critical role, but the max/average/min operation only explores the first-order information, which limits the discrimination ability of the CNN features in some complex situations. In this paper, a high-order pooling layer is integrated into the VGG16 network for visual tracking. In detail, a high-order covariance pooling layer is employed to replace the last maxpooling layer to learn discrimination features and is trained on the ImageNet and CUB200-2011 data sets. In tracking stage, the multiple levels of feature maps are extracted as the appearance representation of the target. After that, the extracted CNN features are integrated into the correlation filters framework when tracking is on-the-fly. The experimental results show that the proposed algorithm achieves excellent performance in both success rate and tracking accuracy.

Printed Circuit Boards Defect Detection Method Based on Improved Fully Convolutional Networks

Since printed circuit board (PCB) is the key to ensure the reliability of electronic equipment. Therefore, defect detection for PCB is a basic and necessary work. This paper proposes a PCB defect detection method based on an improved fully convolutional neural networks to detect four types of defects: spurs, mouse bites, short circuits and open circuits. The improved neural networks increase the receptive field and solve the problems of ordinary atrous convolution information loss of continuity by introducing the continuous atrous convolution module. At the same time, we introduce the improved skip layer after the upsampling layer to get the characteristics of multi-scale fusion, improve the resolution of the image. Due to the fewer levels of the whole networks, MobileNet-V2 networks model is adopted in the improved fully convolutional neural networks, which is compared with ResNet-50 and Vgg-16 networks model. Comparison of accuracy of four kinds of PCB classification defects. Experimental results show that the improved fully convolutional neural networks based on MobileNet-V2 have an average recognition accuracy of 92.86% for four types of PCB defects, which is better than the ResNet-50 and Vgg-16 networks models, verifying the effectiveness of the PCB defect recognition classification. Moreover, experiments are carried out in a real environment to verify the feasibility of the proposed method.

Benchmarking Transfer Learning Strategies in Time-Series Imaging: Recommendations for Analyzing Raw Sensor Data

With the growing availability and complexity of time-series sequences, scalable and robust machine learning approaches are required that overcome the sampling challenge of quantitatively sufficient training data. Following the research trend towards the deep learning-based analysis of time-series encoded as images, this study proposes a time-series imaging workflow that overcomes the challenge of quantitatively limited sensor data across domains (i.e., medicine and engineering). After systematically identifying the three relevant dimensions that affect the performance of the deep learning-based analysis of visualized time-series data, we performed a benchmarking evaluation with a total of 24 unique convolutional neural network models. Following a two-level transfer learning investigation, we reveal that fine-tuning the mid-level features results in the best classification performance. As a result, we present an optimized representation of the VGG16 network, which outperforms previous studies in the field. Our approach is accurate, robust, and manifests internal and external validity. By only using the raw time-series data, our model does not require manual feature engineering, being of high practical relevance. As the post-hoc analysis of our results reveals that our model allows automated extraction of meaningful features based on the trend of the underlying time-series data, our study also adds to explainable artificial intelligence. Furthermore, our proposed workflow reduces the sequence length of the input data while preserving all information. Especially with the hurdle of long-term dependencies in sequential time-series data, we overcome related work’s limitation of the vanishing gradients problem and contribute to the sequential learning theory in artificial intelligence.

Computer Assisted Instruction in Laparoscopic Surgery using Deep Learning

Minimally invasive surgery (MIS) is famous to cause less harm to the skin compared with regular operation, due to the tiny surgical instruments and the small incisions used. It provides many advantages to the patients like a shorter hospital stays, reduced pain and faster recovery. In addition, MIS offers the possibility of video record the surgery. These videos are used for teaching purposes, evaluating surgeons and also they are treated as evidence in case of lawsuits from patients. On the other hand, these types of surgeries are difficult to learn and teach. That’s why surgeons tend to check MIS videos for a possible technical error. Since MIS medias are commonly very long, this manual surgical quality assessment (SQA) process, without any support of video search, take so much time and effort. To surmount this issue, we present a neural network based solution, to identify surgical instruments and index these videos, using three fine-tuned Convolutional Neural Network VGG19, Inception v-4 and NASNet-A. Finally, we present the benefits of the proposed approach on the Cholec80 dataset.

Plant Disease Detection using AI based VGG-16 Model

Agriculture and modern farming is one of the fields where IoT and automation can have a great impact. Maintaining healthy plants and monitoring their environment in order to identify or detect diseases is essential in order to maintain a maximum crop yield. The implementation of current high rocketing technologies including artificial intelligence (AI), machine learning, and deep learning has proved to be extremely important in modern agriculture as a method of advanced image analysis domain. Artificial intelligence adds time efficiency and the possibility of identifying plant diseases, in addition to monitoring and controlling the environmental conditions in farms. Several studies showed that machine learning and deep learning technologies can detect plant diseases upon analyzing plant leaves with great accuracy and sensitivity. In this study, considering the worth of machine learning for disease detection, we present a convolutional neural network VGG-16 model to detect plant diseases, to allow farmers to make timely actions with respect to treatment without further delay. To carry this out, 19 different classes of plants diseases were chosen, where 15,915 plant leaf images (both diseased and healthy leaves) were acquired from the Plant Village dataset for training and testing. Based on the experimental results, the proposed model is able to achieve an accuracy of about 95.2% with the testing loss being only 0.4418. The proposed model provides a clear direction toward a deep learning-based plant disease detection to apply on a large scale in future.

О РАСПОЗНАВАНИИ ОПУХОЛЕЙ ГОЛОВНОГО МОЗГА МЕТОДАМИ ГЛУБОКОГО ОБУЧЕНИЯ

In this paper computer studies of the effectiveness of using transfer learning methods to solve the problem of recognizing human brain tumor based on MRI images were performed. The deep convolutional networks VGG16, ResNet50, InceptionV3, Xception, DenseNet121 and MobileNet_v2 were used as the basic pretrained models. Various training and finetuning strategies for deep convolutional networks for recognizing brain tumor are proposed. An analysis of their performance showed that the strategy of finetuning of the Xception model on an extended MRIscans data set yielded higher accuracy, precision, recall and AUC metrics compared to the other models. The best classification achieved accuracy on the Xception finetuned model is 98%.

Litchi detection in the field using an improved YOLOv3 model

Due to the illumination, complex background, and occlusion of the litchi fruits, the accurate detection of litchi in the field is extremely challenging. In order to solve the problem of the low recognition rate of litchi-picking robots in field conditions, this study was inspired by the ideas of ResNet and dense convolution and proposed an improved feature-extraction network model named “YOLOv3_Litchi”, combining dense connections and residuals for the detection of litchis. Firstly, based on the traditional YOLOv3 deep convolution neural network and regression detection, the idea of residuals was to be put into the feature-extraction network to effectively avoid the problem of decreasing detection accuracy due to the excessive depths of the network layers. Secondly, under the premise of a good receptive field and high detection accuracy, the large convolution kernel was replaced by a small convolution kernel in the shallow layer of the network, thereby effectively reducing the model parameters. Finally, the idea of feature pyramid was used to design the network to identify the small target litchi to ensure that the shallow features were not lost and simultaneously reduced the model parameters. Experimental results show that the improved YOLOv3_Litchi model achieved better results than the classic YOLOv3_DarkNet-53 model and the YOLOv3_Tiny model. The mean average precision (mAP) score was 97.07%, which was higher than the 95.18% mAP of the YOLOv3_DarkNet-53 model and the 94.48% mAP of the YOLOv3_Tiny model. The frame frequency was 58 fps, which was higher than 29 fps of the YOLOv3_DarkNet-53 model. Compared with the classic Faster R-CNN model with the feature-extraction network VGG16, the mAP was increased by 1%, and the FPS advantage was obvious. Compared with the classic single shot multibox detector (SSD) model, both the accuracy and the running efficiency were improved. The results show that the improved YOLOv3_Litchi model had stronger robustness, higher detection accuracy, and less computational complexity for the identification of litchi in the field conditions, which should be helpful for litchi orchard precision management. Keywords: deep learning; residual network; dense connection; feature pyramid network DOI: 10.25165/j.ijabe.20221502.6541 Citation: Peng H X, Xue C, Shao Y Y, Chen K Y, Liu H N, Xiong J T, et al. Litchi detection in the field using an improved YOLOv3 model. Int J Agric & Biol Eng, 2022; 15(2): 211–220.

ACRnet: Adaptive Cross-transfer Residual neural network for chest X-ray images discrimination of the cardiothoracic diseases.

Cardiothoracic diseases are a serious threat to human health and chest X-ray image is a great reference in diagnosis and treatment. At present, it has been a research hot-spot how to recognize chest X-ray image automatically and exactly by the computer vision technology, and many scholars have gotten the excited research achievements. While both emphysema and cardiomegaly often are associated, and the symptom of them are very similar, so the X-ray images discrimination for them led easily to misdiagnosis too. Therefore, some efforts are still expected to develop a higher precision and better performance deep learning model to recognize efficiently the two diseases. In this work, we construct an adaptive cross-transfer residual neural network (ACRnet) to identify emphysema, cardiomegaly and normal. We cross-transfer the information extracted by the residual block and adaptive structure to different levels in ACRnet, and the method avoids the reduction of the adaptive function by residual structure and improves the recognition performance of the model. To evaluate the recognition ability of ACRnet, four neural networks VGG16, InceptionV2, ResNet101 and CliqueNet are used for comparison. The results show that ACRnet has better recognition ability than other networks. In addition, we use the deep convolution generative adversarial network (DCGAN) to expand the original dataset and ACRnet's recognition ability is greatly improved.

Detection and Classification of Leaf Disease Using Deep Learning for a Greenhouses’ Robot

Plant diseases are a severe threat to the environment, economy, and health. Early disease identification remains a challenging task in Iraq due to the scarce of the necessary resources and infrastructure. This paper uses various deep learning algorithms to detect different diseases on plant leaves and detect healthy ones, using an RGB camera as a crucial part of our real-time autonomous greenhouses' robot along with using two datasets, plant-village and cotton dataset, to investigate the best convolutional neural network architecture. The first dataset contained 10,190 images from the plant-village open datasets; it includes four crops with ten distinct classes of diseased and healthy leaves. Moreover, the cotton dataset contained 2,204 images for training and 106 images for testing; it has four classes of diseased and healthy plants and leaves. Different network architectures were tested in this paper for the best suitable lightweight architecture for our mobile robot. Results show that the best performance is 99.908% which achieved by the VGG16 network. The highest accuracy of VGG16 obtained in our research makes it the best tool for our autonomous plant disease detection robot. Index Terms— Convolutional neural networks architectures, Computer vision, Deep learning, Leaf disease detection and classification, Precision agriculture.

Proposing Algorithm Using YOLOV4 and VGG-16 for Smart-Education

In this paper, we propose an algorithm to identify and solve systems of high-order equations. We rely on traditional solution methods to build algorithms to solve automated equations based on deep learning. The proposal method includes two main steps. In the first step, we use YOLOV4 (Kumar et al. 2020; Canu, 2020) to recognize equations and letters associated with the VGG-16 network (Simonyan and Zisserman, 2015) to classify them. We then used the SymPy model to solve the equations in the second step. Data are images of systems of equations that are typed and designed by ourselves or handwritten from other sources. Besides, we also built a web-based application that helps users select an image from their devices. The results show that the proposed algorithm is set out with 95% accuracy for smart-education applications.

Salient object detection based on global to local visual search guidance

The human eye has the behavior of fixation shift, which is represented by visual search mechanism. Human visual search mechanism has the characteristics of visual guidance from global to local. In order to imitate this behavior, this paper proposes a novel salient object detection model which integrates global and local perception based on visual search guidance. First, by using multi-level feature complementarity, a coarse global perceptual map can be obtained by encoding VGG network and decoding deconvolution network. Then, through the Maximally Stable External Regions (MSER) algorithm, the local hotspot sub-regions in the coarse global perceptual map are intercepted, and generating a visual guidance search path based on the maximum response principle, and the local fine perceptual map is successively merged through the Squeeze-and-Excitation Networks (SENet) step by step. Finally, the salient map integrating the coarse global and fine local perception is achieved. Because only one local area needs to be integrated at a time, which is equivalent to solving a relatively simple problem at a time, it is easy to get more accurate results. The experimental results show that the time-consuming of our model is relatively low and compared with the classical and more recent SOTA methods, our model also shows competitive results.

Drone-Based Water Level Detection in Flood Disasters.

Japan was hit by typhoon Hagibis, which came with torrential rains submerging almost eight-thousand buildings. For fast alleviation of and recovery from flood damage, a quick, broad, and accurate assessment of the damage situation is required. Image analysis provides a much more feasible alternative than on-site sensors due to their installation and maintenance costs. Nevertheless, most state-of-art research relies on only ground-level images that are inevitably limited in their field of vision. This paper presents a water level detection system based on aerial drone-based image recognition. The system applies the R-CNN learning model together with a novel labeling method on the reference objects, including houses and cars. The proposed system tackles the challenges of the limited and wild data set of flood images from the top view with data augmentation and transfer-learning overlaying Mask R-CNN for the object recognition model. Additionally, the VGG16 network is employed for water level detection purposes. We evaluated the proposed system on realistic images captured at disaster time. Preliminary results show that the system can achieve a detection accuracy of submerged objects of 73.42% with as low as only 21.43 cm error in estimating the water level.

Expression Recognition Method Using Improved VGG16 Network Model in Robot Interaction

Aiming at the problems of poor representation ability and less feature data when traditional expression recognition methods are applied to intelligent applications, an expression recognition method based on improved VGG16 network is proposed. Firstly, the VGG16 network is improved by using large convolution kernel instead of small convolution kernel and reducing some fully connected layers to reduce the complexity and parameters of the model. Then, the high-dimensional abstract feature data output by the improved VGG16 is input into the convolution neural network (CNN) for training, so as to output the expression types with high accuracy. Finally, the expression recognition method combined with the improved VGG16 and CNN model is applied to the human-computer interaction of the NAO robot. The robot makes different interactive actions according to different expressions. The experimental results based on CK + dataset show that the improved VGG16 network has strong supervised learning ability. It can extract features well for different expression types, and its overall recognition accuracy is close to 90%. Through multiple tests, the interactive results show that the robot can stably recognize emotions and make corresponding action interactions.

Smart Pothole Detection Using Deep Learning Based on Dilated Convolution.

Roads make a huge contribution to the economy and act as a platform for transportation. Potholes in roads are one of the major concerns in transportation infrastructure. A lot of research has proposed using computer vision techniques to automate pothole detection that include a wide range of image processing and object detection algorithms. There is a need to automate the pothole detection process with adequate accuracy and speed and implement the process easily and with low setup cost. In this paper, we have developed efficient deep learning convolution neural networks (CNNs) to detect potholes in real-time with adequate accuracy. To reduce the computational cost and improve the training results, this paper proposes a modified VGG16 (MVGG16) network by removing some convolution layers and using different dilation rates. Moreover, this paper uses the MVGG16 as a backbone network for the Faster R-CNN. In addition, this work compares the performance of YOLOv5 (Large (Yl), Medium (Ym), and Small (Ys)) models with ResNet101 backbone and Faster R-CNN with ResNet50(FPN), VGG16, MobileNetV2, InceptionV3, and MVGG16 backbones. The experimental results show that the Ys model is more applicable for real-time pothole detection because of its speed. In addition, using the MVGG16 network as the backbone of the Faster R-CNN provides better mean precision and shorter inference time than using VGG16, InceptionV3, or MobilNetV2 backbones. The proposed MVGG16 succeeds in balancing the pothole detection accuracy and speed.

Characterizing Malignant Melanoma Clinically Resembling Seborrheic Keratosis Using Deep Knowledge Transfer.

Simple SummaryMalignant melanomas (MMs) with aypical clinical presentation constitute a diagnostic pitfall, and false negatives carry the risk of a diagnostic delay and improper disease management. Among the most common, challenging presentation forms of MMs are those that clinically resemble seborrheic keratosis (SK). On the other hand, SK may mimic melanoma, producing ‘false positive overdiagnosis’ and leading to needless excisions. The evolving efficiency of deep learning algorithms in image recognition and the availability of large image databases have accelerated the development of advanced computer-aided systems for melanoma detection. In the present study, we used image data from the International Skin Image Collaboration archive to explore the capacity of deep knowledge transfer in the challenging diagnostic task of the atypical skin tumors of MM and SK.Malignant melanomas resembling seborrheic keratosis (SK-like MMs) are atypical, challenging to diagnose melanoma cases that carry the risk of delayed diagnosis and inadequate treatment. On the other hand, SK may mimic melanoma, producing a ‘false positive’ with unnecessary lesion excisions. The present study proposes a computer-based approach using dermoscopy images for the characterization of SΚ-like MMs. Dermoscopic images were retrieved from the International Skin Imaging Collaboration archive. Exploiting image embeddings from pretrained convolutional network VGG16, we trained a support vector machine (SVM) classification model on a data set of 667 images. SVM optimal hyperparameter selection was carried out using the Bayesian optimization method. The classifier was tested on an independent data set of 311 images with atypical appearance: MMs had an absence of pigmented network and had an existence of milia-like cysts. SK lacked milia-like cysts and had a pigmented network. Atypical MMs were characterized with a sensitivity and specificity of 78.6% and 84.5%, respectively. The advent of deep learning in image recognition has attracted the interest of computer science towards improved skin lesion diagnosis. Open-source, public access archives of skin images empower further the implementation and validation of computer-based systems that might contribute significantly to complex clinical diagnostic problems such as the characterization of SK-like MMs.

Deep learning methods in the diagnosis of sacroiliitis from plain pelvic radiographs.

The aim of this study is to develop a computer-aided diagnosis method to assist physicians in evaluating sacroiliac radiographs. Convolutional neural networks, a deep learning method, were used in this retrospective study. Transfer learning was implemented with pre-trained VGG-16, ResNet-101 and Inception-v3 networks. Normal pelvic radiographs (n = 290) and pelvic radiographs with sacroiliitis (n = 295) were used for the training of networks. The training results were evaluated with the criteria of accuracy, sensitivity, specificity and precision calculated from the confusion matrix and AUC (area under the ROC curve) calculated from ROC (receiver operating characteristic) curve. Pre-trained VGG-16 model revealed accuracy, sensitivity, specificity, precision and AUC figures of 89.9%, 90.9%, 88.9%, 88.9% and 0.96 with test images, respectively. These results were 84.3%, 91.9%, 78.8%, 75.6 and 0.92 with pre-trained ResNet-101, and 82.0%, 79.6%, 85.0%, 86.7% and 0.90 with pre-trained inception-v3, respectively. Successful results were obtained with all three models in this study where transfer learning was applied with pre-trained VGG-16, ResNet-101 and Inception-v3 networks. This method can assist clinicians in the diagnosis of sacroiliitis, provide them with a second objective interpretation and also reduce the need for advanced imaging methods such as magnetic resonance imaging.

Metallic debossed characters industrial online non‐segmentation identification based on improved multi‐scale image fusion enhancement and deep neural network

This paper investigates accurate industry online scene text recognition techniques for metallic debossed characters (MDCs). As MDCs have low contrast with its background and easy to interfere with corrosion rust, oxidate skin or specular reflection and so forth, a multi-scale image fusion algorithm is first proposed to restore the MDCs’ shape and appearance information to enhance the sample's contrast by using the MDCs' depth characteristic with the help of four-directional illumination. With this method, the three most important reflection-related picture quality evaluation standards, namely contrast, saturation, and exposure, can be enhanced simultaneously. Next, based on the fusion result, a U-shaped network based on VGG-16 network architecture is used for text localization. In order to improve network receptive fields, an improved post-processing output module that combines a text sequence score map for coarsening localization with a single-character score map for fine localization is proposed, so that the network's adaptability and accuracy for long texts are improved. Finally, a convolutional recurrent neural network is adopted to realize the recognition of sequence characters. The advantages and effectiveness of the proposed method are statistically analyzed with the data from a liquefied petroleum gas cylinder annual periodic inspection production line.