Deep Convolutional Neural Network Method Research Articles

A Method of License Plate Location and Character Recognition based on CNN

At the present time, the economy continues to flourish, and private cars have become the means of choice for most people. Therefore, the license plate recognition technology has become an indispensable part of intelligent transportation, with research and application value. In recent years，the convolution neural network for image classification is an application of deep learning on image processing. This paper proposes a strategy to improve the YOLO model by studying the deep learning convolutional neural network (CNN) and related target detection methods, and combines the OpenCV and TensorFlow frameworks to achieve efficient recognition of license plate characters. The experimental results show that target detection method based on YOLO is beneficial to shorten the training process and achieve a good level of accuracy.

Autonomous Fruit Recognition System Based on Deep Convolutional Neural Network

Recently it is found that people are becoming more cautious to their diet throughout the universe. Unhealthy diet can cause many problems like sugar, obesity, gain in weight and many other chronic health related issues. Essential part of our diet is contributed by fruits as they are rich source of vitamins, fiber, energy and nutrients. Today’s era has been adapted to a system of intake of foods which has several adverse effects on human health. The proposed system is Autonomous Fruit Recognition system based on Deep Convolutional Neural Network (DCNN) method. Using this technology recognition and estimation of fruit calories is necessary to spread awareness about food habits among people suffering from obesity due to bad food culture and consumption of food .This proposed web/app based system simplifies the calorie measuring process of fruit. The machine learning based API used in our system recognize the fruit and provide calorie content of that fruit. System uses convolutional Neural Network called MobileNet. This web/app based application is user friendly.

Bearing fault diagnosis based on improved VMD and DCNN

Vibration signal produced by rolling element bearings has obvious non-stationary and nonlinear characteristics, and it’s necessary to preprocess the original signals to obtain better diagnostic results. This paper proposes an improved variational mode decomposition (IVMD) and deep convolutional neural network (DCNN) method to realize the intelligent fault diagnosis of rolling element bearings. Firstly, to solve the problem that the number of decomposed modes of variational mode decomposition (VMD) needs to be preset, an IVMD method is proposed, where the mode number can be determined adaptively according to the curve of the instantaneous frequency mean of mode functions. With this method, the vibration signal can be decomposed into a series of modal components containing bearing fault characteristic information. Then, DCNN is employed to fuse these multi-scale modal components, which can automatically learn fault features and establish bearing fault diagnosis model to realize intelligent fault diagnosis eventually. Experimental analysis and comparison results verify that the proposed method can effectively enhance the bearing fault features and improve the diagnosis accuracy.

Handwriting identification using deep convolutional neural network method

Handwriting is a unique thing that produced differently for each person. Handwriting has a characteristic that remain the same with single writer, so a handwriting can be used as a variable in biometric systems. Each person have a different form of handwriting style but with a small possibility that same characters have something commons. This paper proposes a handwriting identification method using sentence segmented handwriting forms. Sentence form is used to get more complete handwriting characteristics than using a single characters or words. Dataset used is divided into three categories of images, binary, grayscale, and inverted binary. All datasets have same image with different in color and consist of 100 class. Transfer learning used in this paper are pre-trained model VGG19. Training was conducted in 100 epochs. Highest result is grayscale images with genuince acceptance rate of 92.3% and equal error rate of 7.7%.

Detection of Mesangial hypercellularity of MEST-C score in immunoglobulin A-nephropathy using deep convolutional neural network

Immunoglobulin A (IgA)-nephropathy (IgAN) is one of the major reasons for renal failure. It provides vital clues to estimate the stage and the proliferation rate of end-stage kidney disease. IgA stage can be estimated with the help of MEST-C score. The manual estimation of MEST-C score from whole slide kidney images is a very tedious and difficult task. This study uses some Convolutional neural networks (CNNs) related models to detect mesangial hypercellularity (M score) in MEST-C. CNN learns the features directly from image data without the requirement of analytical data. CNN is trained efficiently when image data size is large enough for a particular class. In the case of smaller data size, transfer learning can be used efficiently in which CNN is pre-trained on some general images and then on subject images. Since the data set size is small, time spent in collecting large data set is saved. The training time of transfer learning is also reduced because the model is already pre-trained. This research work aims at the detection of mesangial hypercellularity from biopsy images with small data size by utilizing the transfer learning. The dataset used in this research work consists of 138 individual glomerulus (× 20 magnification digital biopsy) images of IgA patients received from All India Institute of Medical Science, Delhi. Here, machine learning (k-nearest neighbour (KNN) and support vector machine (SVM)) classifiers are compared to transfer learning CNN methods. The deep extracted image features are used by machine learning classifiers. The different evaluation parameters have been used for comparing the predictions of basic classifiers to the deep learning model. The research work concludes that the transfer learning deep CNN method can improve the detection of mesangial hypercellularity as compare to KNN, SVM methods when using the small data set. This model could help the pathologists to understand the stages of kidney failure.

Automated detection and delineation of hepatocellular carcinoma on multiphasic contrast-enhanced MRI using deep learning.

Liver Imaging Reporting and Data System (LI-RADS) uses multiphasic contrast-enhanced imaging for hepatocellular carcinoma (HCC) diagnosis. The goal of this feasibility study was to establish a proof-of-principle concept towards automating the application of LI-RADS, using a deep learning algorithm trained to segment the liver and delineate HCCs on MRI automatically. In this retrospective single-center study, multiphasic contrast-enhanced MRIs using T1-weighted breath-hold sequences acquired from 2010 to 2018 were used to train a deep convolutional neural network (DCNN) with a U-Net architecture. The U-Net was trained (using 70% of all data), validated (15%) and tested (15%) on 174 patients with 231 lesions. Manual 3D segmentations of the liver and HCC were ground truth. The dice similarity coefficient (DSC) was measured between manual and DCNN methods. Postprocessing using a random forest (RF) classifier employing radiomic features and thresholding (TR) of the mean neural activation was used to reduce the average false positive rate (AFPR). 73 and 75% of HCCs were detected on validation and test sets, respectively, using > 0.2 DSC criterion between individual lesions and their corresponding segmentations. Validation set AFPRs were 2.81, 0.77, 0.85 for U-Net, U-Net + RF, and U-Net + TR, respectively. Combining both RF and TR with the U-Net improved the AFPR to 0.62 and 0.75 for the validation and test sets, respectively. Mean DSC between automatically detected lesions using the DCNN + RF + TR and corresponding manual segmentations was 0.64/0.68 (validation/test), and 0.91/0.91 for liver segmentations. Our DCNN approach can segment the liver and HCCs automatically. This could enable a more workflow efficient and clinically realistic implementation of LI-RADS.

A survey on deep learning methods for scene flow estimation

Recently, computer vision has achieved remarkable accomplishments in many domains under the thriving of deep learning. Scene flow estimation turns from the classical manual feature construction to the deep convolutional neural network (DCNN) approaches. In this paper, we review recent works about scene flow, mainly focusing on DCNN methods. We present some milestones of scene flow in recent years, and categorize these methods into supervised and unsupervised based methods. Meanwhile, we also review some multi-task methods related to scene flow. At last, we present a performance comparison among different methods.

Deep salient wood image-based quality assessment

We introduce a novelty in the method of the deep salient wood image quality assessment (DS-WIQA) for no-reference image quality assessment (NR-IQA). We exploit a five-layer deep convolutional neural network (DCNN) for the salient wood image map. DS-WIQA also employs the n-convex-concave model. The outcomes obviously prove that our DCNN and DS-WIQA architectures can deliver a superior achievement on Zenodo and Lignoindo datasets, respectively. We compute a salient wood image map of each wood image in small wood image patches. Our exploratory outcomes evince that the proposed DCNN and DS-WIQA methods are superior to other the advanced methods on Zenodo and Lignoindo datasets, respectively. Our proposed DCNN for NR-IQA also obtains a better result compared with the other NR-IQA methods in the five distortion types of JP2K, JPEG, white noise Gaussian, blocking artifact, and the fast fading and also in the undistorted wood images. Our DCNN outruns the recent most sophisticated methods in terms of SROCC and LCC evaluation, respectively. DS-WIQA outpaces other the advanced methods by $$0.38\%$$ and $$0.22\%$$ greater than our proposed DCNN, and $$34.84\%$$ and $$30.15\%$$ greater than other methods with respect to SROCC and LCC, respectively. In computational complexity of our proposed DCNN and DS-WIQA cut down the shift add operation in exponential, logarithmic, and trigonometric functions. DS-WIQA shows up to be more significant than our proposed DCNN and the other DCNN methods.

Fault Diagnosis under Variable Working Conditions Based on STFT and Transfer Deep Residual Network

Fault diagnosis plays a very important role in ensuring the safe and reliable operations of machines. Currently, the deep learning-based fault diagnosis is attracting increasing attention. However, fault diagnosis under variable working conditions has been a significant challenge due to the domain discrepancy problem. This problem is also unavoidable in deep learning-based fault diagnosis methods. This paper contributes to the ongoing investigation by proposing a new approach for the fault diagnosis under variable working conditions based on STFT and transfer deep residual network (TDRN). The STFT was employed to convert vibration signal to time-frequency image as the input of the TDRN. To address the domain discrepancy problem, the TDRN was developed in this paper. Unlike traditional deep convolutional neural network (DCNN) methods, by combining with transfer learning, the TDRN can make a bridge between two different working conditions, thereby using the knowledge learned from a working condition to achieve a high classification accuracy in another working condition. Moreover, since the residual learning is introducing, the TDRN can overcome the problems of training difficulty and performance degradation existing in traditional DCNN methods, thus further improving the classification accuracy. Experiments are conducted on the popular CWRU bearing dataset to validate the effectiveness and superiority of the proposed approach. The results show that the developed TDRN outperforms those methods without transfer learning and/or residual learning in terms of the accuracy and feature learning ability for the fault diagnosis under variable working conditions.

Autonomous Fruit Recognition System based on Deep Convolutional Neural Network

Recently it is found that people are becoming more cautious to their diet throughout the universe. Unhealthy diet can cause many problems like sugar, obesity, gain in weight and many other chronic health related issues. Essential part of our diet is contributed by fruits as they are rich source of vitamins,fiber,energy and nutrients. Today's era has been adapted to a system of intake of foods which has several adverse effects on human health. The proposed system is Autonomous Fruit Recognition system based on Deep Convolutional Neural Network (DCNN) method. Using this technology recognition and estimation of fruit calories is necessary to spread awareness about food habits among people suffering from obesity due to bad food culture and consumption of food .This proposed web/app based system simplifies the calorie measuring process of fruit. The machine learning based API used in our system recognize the fruit and provide calorie content of that fruit. System uses convolutional Neural Network called MobileNet. This web/app based application is user friendly.

Sentiment analysis of customer response of telecommunication operator in Twitter using DCNN-SVM Algorithm

Along with the development of the times, social media is in great demand by various circles of society because social media allows users to express their thoughts or feelings freely. It is important for a company to know public responses about the product or service offered. With this public response, companies can analyze customer needs and plan more satisfying products or services. To be able to know the sentiments of responses, it is necessary to classify responses. Therefore, in this study used the Deep Convolutional Neural Network (DCNN) method as a feature extraction and Support Vector Machine (SVM) as its classification. The performance results of this research are 63% for accuracy, 63% for precision, and 50% for recall of test data.

Marine Organism Detection and Classification from Underwater Vision Based on the Deep CNN Method

Seabed fishing depends on humans in common, for instance, the sea cucumber, sea urchin, and scallop fishing, which is always a very dangerous task. Considering the underwater complex environment conditions such as low temperature, dim vision, and high pressure, collecting the marine products using underwater robots is commonly regarded as a feasible solution. The key technique of the underwater robot development is to detect and locate the main target from underwater vision. This research is based on the deep convolutional neural network (CNN) to realize the target recognition from underwater vision. The RPN (Region Proposal Network) is used to optimize the feature extraction capability. Deep learning dataset is prepared using an underwater video obtained from a sea cucumber fishing ROV (Remote Operated Vehicle). The inspiration of the network structure and the improvements come from the Faster RCNN and Hypernet method, and for the underwater dataset, the method proposed in this paper shows a good performance of recall and object detection accuracy. The detection runs with a speed of 17 fps on a GPU, which is applicable to be used for real-time processing.

Evaluation of artificial intelligence for detecting periapical pathosis on cone‐beam computed tomography scans

To verify the diagnostic performance of an artificial intelligence system based on the deep convolutional neural network method to detect periapical pathosis on cone-beam computed tomography (CBCT) images. images of 153 periapical lesions obtained from 109 patients were included. The specific area of the jaw and teeth associated with the periapical lesions were then determined by a human observer. Lesion volumes were calculated using the manual segmentation methods using Fujifilm-Synapse 3D software (Fujifilm Medical Systems, Tokyo, Japan). The neural network was then used to determine (i) whether the lesion could be detected; (ii) if the lesion was detected, where it was localized (maxilla, mandible or specific tooth); and (iii) lesion volume. Manual segmentation and artificial intelligence (AI) (Diagnocat Inc., San Francisco, CA, USA) methods were compared using Wilcoxon signed rank test and Bland-Altman analysis. The deep convolutional neural network system was successful in detecting teeth and numbering specific teeth. Only one tooth was incorrectly identified. The AI system was able to detect 142 of a total of 153 periapical lesions. The reliability of correctly detecting a periapical lesion was 92.8%. The deep convolutional neural network volumetric measurements of the lesions were similar to those with manual segmentation. There was no significant difference between the two measurement methods (P>0.05). Volume measurements performed by humans and by AI systems were comparable to each other. AI systems based on deep learning methods can be useful for detecting periapical pathosis on CBCT images for clinical application.

Discriminative dictionary learning based sparse representation classification for intelligent fault identification of planet bearings in wind turbine

Planet bearing fault identification is an attractive but challenging task in wind turbine condition monitoring and fault diagnosis. Traditional fault characteristic frequency identification based diagnostic strategies are not sufficient for reliable planet bearing fault detection, due to complex physical configurations and modulation characteristics in planetary drivetrains. In this paper, we propose a discriminative dictionary learning based sparse representation classification (DDL-SRC) framework for intelligent planet bearing fault identification. Our framework could learn a reconstructive and discriminative dictionary for signal sparse representation and an optimal linear classifier for classification tasks simultaneously, which bridges the gap between dictionary learning and classifier training in traditional SRC methods. Specifically, the optimization objective introduces a novel regularization term called ‘discriminative sparse codes error’ and incorporates it with the reconstruction error and classification error. Thus, the dictionary learned by our framework possesses not only the reconstructive power for sparse representation but also the discriminative power for classifier training. The optimization formulation is efficiently solved using K-SVD and orthogonal matching pursuit algorithms. The experiment validations have been conducted for demonstrating the effectiveness and superiority of the proposed DDL-SRC framework over the state-of-the-art dictionary learning based SRC and deep convolutional neural network methods for intelligent planet bearing fault identification.

Sea-Land Segmentation Using Deep Learning Techniques for Landsat-8 OLI Imagery

Automated coastline extraction from optical satellites is fundamental to coastal mapping, and sea-land segmentation is the core technology of coastline extraction. Deep convolutional neural networks (DCNNs) have performed well in semantic segmentation in recent years. However, sea-land segmentation using deep learning techniques remains a challenging task, due to the lack of a benchmark dataset and the difficulty of deciding which semantic segmentation model to use. We present a comparative framework of sea-land segmentation to Landsat-8 OLI imagery via semantic segmentation in deep learning techniques. Three issues are investigated: (1) constructing a sea-land benchmark dataset using Landsat-8 Operational Land Imager (OLI) imagery consisting of 18,000 km2 of coastline around China; (2) evaluating the feasibility and performance of sea-land segmentation by comparing the accuracy assessment, time complexity, spatial complexity and stability of state-of-the-art DCNNs methods; (3) choosing the most suitable semantic segmentation model for sea-land segmentation in accordance with Akaike information criterion (AIC) and Bayesian information criterion (BIC) model selection. Results show that the average test accuracy achieves over 99% accuracy, and the mean Intersection over Unions (mean IoU) is above 92%. These findings demonstrate that the Fully Convolutional DenseNet (FC-DenseNet) performs better than other state-of-the-art methods in sea-land segmentation, based on both AIC and BIC. Considering training time efficiency, DeeplabV3+ performs better for sea-land segmentation. The sea-land segmentation benchmark dataset is available at: https://pan.baidu.com/s/1BlnHiltOLbLKe4TG8lZ5xg.

Optimal Feature Search for Vigilance Estimation Using Deep Reinforcement Learning

A low level of vigilance is one of the main reasons for traffic and industrial accidents. We conducted experiments to evoke the low level of vigilance and record physiological data through single-channel electroencephalogram (EEG) and electrocardiogram (ECG) measurements. In this study, a deep Q-network (DQN) algorithm was designed, using conventional feature engineering and deep convolutional neural network (CNN) methods, to extract the optimal features. The DQN yielded the optimal features: two CNN features from ECG and two conventional features from EEG. The ECG features were more significant for tracking the transitions within the alertness continuum with the DQN. The classification was performed with a small number of features, and the results were similar to those from using all of the features. This suggests that the DQN could be applied to investigating biomarkers for physiological responses and optimizing the classification system to reduce the input resources.

DCNN With Explicable Training Guide and its Application to Fault Diagnosis of the Planetary Gearboxes

The diagnosis performance of Deep Convolutional Neural Network (DCNN) method is closely related to the generalization ability of the training model. An empirical training strategy is to randomly disperse the training samples and train the model with mini-batch training samples. But there are still two problems in the empirical method that need to be solved urgently. Firstly, what is the theoretical basis for random discretization of samples? Secondly, how to scientifically quantify batch division? Aiming at these two problems, the theoretical basis of sample random discretization has been deduced and proved, furthermore, a scientific quantitative batch division method is proposed based on the proved thesis. The fault diagnosis results of the planetary gearbox show that: (1) The model obtained by the training guide proposed in this paper has stronger generalization ability; (2) The DCNN with the training guide can accurately and effectively diagnose the faults of planetary gearbox and obtain ideal diagnosis results.

Deep Convolutional Network for Stereo Depth Mapping in Binocular Endoscopy

Depth mapping from binocular endoscopy images plays an important role in stereoscopic surgical treatment. Owing to the development of deep convolutional neural networks (CNNs), binocular depth estimation models have achieved many exciting results in the fields of autonomous driving and machine vision. However, the application of these methods to endoscopic imaging is greatly limited by the fact that binocular endoscopic images not only are rare, but also have unsatisfying features such as no texture, no ground truth, bad contrast, and high gloss. Aiming at solving the above-mentioned problems, we have built a precise gastrointestinal environment by the open-source software blender to simulate abundant binocular endoscopy data and proposed a 23-layer deep CNNs method to generate real-time stereo depth mapping. An efficient scale-invariant loss function is introduced in this paper to accommodate the characteristics of endoscope images, which improves the accuracy of achieved depth mapping results. Regarding the considerable training data for typical CNNs, our method requires only a few images ($960\times 720$ resolution) at 45 frames per second on an NVIDIA GTX 1080 GPU module, then the depth mapping information is generated in real-time with satisfactory accuracy. The effectiveness of the developed method is validated by comparing with state-of-the-art methods on processing the same datasets, demonstrating a faster and more accurate performance than other model frames.

Seismic horizon tracking using a deep convolutional neural network

Seismic horizons are essential for structural analysis, inversion, time-to-depth conversion, and seismic attribution analysis. However, seismic horizons are obtained commonly by manual tracking or a combination of manual tracking and traditional auto-tracking techniques, which is a time consuming and error-prone process. Although many conventional auto-tracking techniques have been proposed to improve the efficiency of horizon tracking, there are still many challenges to track seismic horizons with complex seismic reflection signatures. A deep convolutional neural network (CNN) method, a typical kind of deep neural networks in the field of deep learning, was proposed in this paper to solve the problems. Seismic horizon tracking is formulated as an image classification task. CNN can take full advantage of the high flexibility in network architecture and hierarchical feature extraction ability to track horizons from seismic volumes in which discontinuities, lateral variations and faults are apparent. We designed a CNN structure for horizon tracking and trained CNN models by using only 1% and 1.5% of the seismic volume to perform a single horizon and multiple horizon tracking. The root mean square error (RMSE) and the coefficient of determination (R2) between the horizon tracked by CNN model and the manual-tracking are 1.78 and 0.998 in single horizon tracking, respectively. The average RMSE and R2 are 2.68 and 0.992 in multiple horizon tracking, respectively. The research we have done suggests that the proposed method can track seismic horizons from 3-D seismic volumes much more accurately and efficiently than conventional methods.

Performance Analysis and Dynamic Evolution of Deep Convolutional Neural Network for Electromagnetic Inverse Scattering

The solution of nonlinear electromagnetic (EM) inverse scattering problems is typically hindered by several challenges such as ill-posedness, strong nonlinearity, and high computational costs. Recently, deep learning has been demonstrated to be a promising tool in addressing these challenges. In particular, it is possible to establish a connection between a deep convolutional neural network (CNN) and iterative solution methods of nonlinear EM inverse scattering. This has led to the development of an efficient CNN-based solution to nonlinear EM inverse problems, termed DeepNIS. It has been shown that DeepNIS can outperform conventional nonlinear inverse scattering methods in terms of both image quality and computational time. In this work, we quantitatively evaluate the performance of DeepNIS as a function of the number of layers using structure similarity measure (SSIM) and mean-square error (MSE) metrics. In addition, we probe the dynamic evolution behavior of DeepNIS by examining its near-isometry property. It is shown that after a proper training stage the proposed CNN is near optimal in terms of the stability and generalization ability.