Learning-based Convolutional Neural Network Research Articles

Classification and Identification of Diabetic Retinopathy Severity Stages in Thai Patients Using Deep Learning Based Convolution Neural Network

Classification and Identification of Diabetic Retinopathy Severity Stages in Thai Patients Using Deep Learning Based Convolution Neural Network

Detection of dental diseases from radiographic 2d dental image using hybrid graph-cut technique and convolutional neural network

In the present scenario the major oral health issues of man is crucial an area of research. The Data mining techniques, image processing, and Computational intelligence techniques are playing a vital role in biomedical research. Dental image processing helps to improve the early detection and classification of the diagnostic process to make accurate decisions. The radiographic 2d dental image is widely utilized for analytic thinking of several dental disorders. In this paper traces the complete steps such as classification and segmentation as well as pre-processing of dental images has been carried out. In the pre-processing, histogram based on adaptive approach is used to stretch the contrast and equalize the brightness throughout the radiographic X-ray 2d dental Image. This operation is useful to distinguish the foreground teeth and the regions of background bones. Separation of dental 2d images into regions corresponding to the objects is a fundamental step of segmentation. The hybrid graph cut segmentation is used to segment the oral cavity and its tissues. In this research deep learning based convolution neural network (CNN) has been used to process the dental image and shows promising outcomes with 97.07% accuracy.

PoreNet: CNN-based Pore Descriptor for High-resolution Fingerprint Recognition

With the development of high-resolution fingerprint scanners, high-resolution fingerprint-based biometric recognition has received increasing attention in recent years. This paper presents a pore feature-based approach for biometric recognition. Our approach employs a convolutional neural network (CNN) model, DeepResPore, to detect pores in the input fingerprint image. Thereafter, a CNN-based descriptor is computed for a patch around each detected pore. Specifically, we have designed a residual learning-based CNN, referred to as PoreNet that learns distinctive feature representation from pore patches. For verification, the match score is generated by comparing pore descriptors obtained from a pair of fingerprint images in bi-directional manner using the Euclidean distance. The proposed approach for high-resolution fingerprint recognition achieves 2.91% and 0.57% equal error rates (EERs) on partial (DBI) and complete (DBII) fingerprints of the benchmark PolyU HRF dataset. Most importantly, it achieves lower FMR1000 and FMR10000 values than the current state-of-the-art approach on both the datasets.

A convolutional neural network algorithm for automatic segmentation of head and neck organs at risk using deep lifelong learning.

This study suggests a lifelong learning-based convolutional neural network (LL-CNN) algorithm as a superior alternative to single-task learning approaches for automatic segmentation of head and neck (OARs) organs at risk. Lifelong learning-based convolutional neural network was trained on twelve head and neck OARs simultaneously using a multitask learning framework. Once the weights of the shared network were established, the final multitask convolutional layer was replaced by a single-task convolutional layer. The single-task transfer learning network was trained on each OAR separately with early stoppage. The accuracy of LL-CNN was assessed based on Dice score and root-mean-square error (RMSE) compared to manually delineated contours set as the gold standard. LL-CNN was compared with 2D-UNet, 3D-UNet, a single-task CNN (ST-CNN), and a pure multitask CNN (MT-CNN). Training, validation, and testing followed Kaggle competition rules, where 160 patients were used for training, 20 were used for internal validation, and 20 in a separate test set were used to report final prediction accuracies. On average contours generated with LL-CNN had higher Dice coefficients and lower RMSE than 2D-UNet, 3D-Unet, ST- CNN, and MT-CNN. LL-CNN required ~72hrs to train using a distributed learning framework on 2 Nvidia 1080Ti graphics processing units. LL-CNN required 20s to predict all 12 OARs, which was approximately as fast as the fastest alternative methods with the exception of MT-CNN. This study demonstrated that for head and neck organs at risk, LL-CNN achieves a prediction accuracy superior to all alternative algorithms.

Pore detection in high-resolution fingerprint images using deep residual network

This letter presents a residual learning-based convolutional neural network, referred to as DeepResPore, for detection of pores in high-resolution fingerprint images. Specifically, the proposed DeepResPore model generates a pore intensity map from the input fingerprint image. Subsequently, the local maxima filter is operated on the pore intensity map to identify the pore coordinates. The results of our experiments indicate that the proposed approach is effective in extracting pores with a true detection rate of 94:49% on Test set I and 93:78% on Test set II of the publicly available PolyU HRF dataset. Most importantly, the proposed approach achieves state-of-the-art performance on both test sets.

Deep Learning From Multiple Crowds: A Case Study of Humanitarian Mapping

Satellite images are widely applied in humanitarian mapping that labels buildings, roads, and so on for humanitarian aid and economic development. However, the labeling now is mostly done by volunteers. In this paper, we utilize deep learning to solve humanitarian mapping tasks of a mobile software named MapSwipe. The current deep learning techniques, e.g., convolutional neural network (CNN), can recognize ground objects from satellite images but rely on numerous labels for training for each specific task. We solve this problem by fusing multiple freely accessible crowdsourced geographic data and propose an active learning-based CNN training framework named MC-CNN to deal with the quality issues of the labels extracted from these data, including incompleteness (e.g., some kinds of object are not labeled) and heterogeneity (e.g., different spatial granularities). The method is evaluated with building mapping in South Malawi and road mapping in Guinea with level-18 satellite images provided by Bing Map and volunteered geographic information from OpenStreetMap, MapSwipe, and OsmAnd. The results based on multiple metrics, including Precision, Recall, F1 Score, and area under the receiver operating characteristic curve, show that MC-CNN can fuse the crowdsourced labels for higher prediction performance and be successfully applied in MapSwipe for humanitarian mapping with 85% labor saved and an overall accuracy of 0.86 achieved.

Fusion Image Based Radar Signal Feature Extraction and Modulation Recognition

The development of cognitive radio and radar electronic reconnaissance has put forward an important demand for improving the recognition ability of modulated signals in complex electromagnetic environment. In this paper, we propose a valid radar signal modulation recognition technology under low signal-to-noise ratio (SNR). The recognition technology can recognize 12 different modulation signals, including Costas, LFM, NLFM, BPSK, P1–P4, and T1–T4 codes. First, we propose the image fusion algorithm of non-multi-scale decomposition to fuse images of a single signal with different time–frequency (T–F) methods. Specifically, weights are designed by the principal component analysis, which could combine significative details of T–F images. Second, we adopt transfer learning-based convolutional neural networks and self-training-based stacked autoencoder, which extract the effective information on fusion image, furthermore guarantee the recognition performance. Moreover, multi-feature fusion algorithm is used to fuse features, which reduces redundant information on features and enhances computing efficiency. Finally, the classifier is performed by a classical algorithm called support vector machine. Simulation results show that the average recognition success rate is 95.5% at SNR of −6dB. It is testified that proposed recognition technology possesses good robustness and superiority in RSR with a wide range of SNR.

An unsupervised convolutional neural network-based algorithm for deformable image registration

The purpose of the work is to develop a deep unsupervised learning strategy for cone-beam CT (CBCT) to CT deformable image registration (DIR). This technique uses a deep convolutional inverse graphics network (DCIGN) based DIR algorithm implemented on 2 Nvidia 1080 Ti graphics processing units. The model is comprised of an encoding and decoding stage. The fully-convolutional encoding stage learns hierarchical features and simultaneously forms an information bottleneck, while the decoding stage restores the original dimensionality of the input image. Activations from the encoding stage are used as the input channels to a sparse DIR algorithm. DCIGN was trained using a distributive learning-based convolutional neural network architecture and used 285 head and neck patients to train, validate, and test the algorithm. The accuracy of the DCIGN algorithm was evaluated on 100 synthetic cases and 12 hold out test patient cases. The results indicate that DCIGN performed better than rigid registration, intensity corrected Demons, and landmark-guided deformable image registration for all evaluation metrics. DCIGN required ~14 h to train, and ~3.5 s to make a prediction on a 512 × 512 × 120 voxel image. In conclusion, DCIGN is able to maintain high accuracy in the presence of CBCT noise contamination, while simultaneously preserving high computational efficiency.

Fine-Grained Face Annotation Using Deep Multi-Task CNN.

We present a multi-task learning-based convolutional neural network (MTL-CNN) able to estimate multiple tags describing face images simultaneously. In total, the model is able to estimate up to 74 different face attributes belonging to three distinct recognition tasks: age group, gender and visual attributes (such as hair color, face shape and the presence of makeup). The proposed model shares all the CNN’s parameters among tasks and deals with task-specific estimation through the introduction of two components: (i) a gating mechanism to control activations’ sharing and to adaptively route them across different face attributes; (ii) a module to post-process the predictions in order to take into account the correlation among face attributes. The model is trained by fusing multiple databases for increasing the number of face attributes that can be estimated and using a center loss for disentangling representations among face attributes in the embedding space. Extensive experiments validate the effectiveness of the proposed approach.

Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm

ObjectivesDeep convolutional neural networks (CNNs) are a rapidly emerging new area of medical research, and have yielded impressive results in diagnosis and prediction in the fields of radiology and pathology. The aim of the current study was to evaluate the efficacy of deep CNN algorithms for detection and diagnosis of dental caries on periapical radiographs. Materials and methodsA total of 3000 periapical radiographic images were divided into a training and validation dataset (n = 2400 [80%]) and a test dataset (n = 600 [20%]). A pre-trained GoogLeNet Inception v3 CNN network was used for preprocessing and transfer learning. The diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, and area under the curve (AUC) were calculated for detection and diagnostic performance of the deep CNN algorithm. ResultsThe diagnostic accuracies of premolar, molar, and both premolar and molar models were 89.0% (80.4–93.3), 88.0% (79.2–93.1), and 82.0% (75.5–87.1), respectively. The deep CNN algorithm achieved an AUC of 0.917 (95% CI 0.860–0.975) on premolar, an AUC of 0.890 (95% CI 0.819–0.961) on molar, and an AUC of 0.845 (95% CI 0.790–0.901) on both premolar and molar models. The premolar model provided the best AUC, which was significantly greater than those for other models (P < 0.001). ConclusionsThis study highlighted the potential utility of deep CNN architecture for the detection and diagnosis of dental caries. A deep CNN algorithm provided considerably good performance in detecting dental caries in periapical radiographs. Clinical significanceDeep CNN algorithms are expected to be among the most effective and efficient methods for diagnosing dental caries.

SURFACE GRADING USING DEEP LEARNING APPROACH FOR CERAMIC TILES

The surface grading of ceramic tiles is essential for ceramic tile industries due to the huge development of infrastructure and essential usage of ceramic tiles. In some industries, surface grading is performed manually. It is a difficult task due to a large number of variations in the surface properties. In this study, a technique for surface grading of ceramic using deep learning is presented. The system uses the VxC Tiles of Surface Grading (TSG) database for performance evaluation. The deep learning based Convolution Neural Network (CNN) is used for the surface grading approach that classifies the tiles into Grade-1 (G1), Grade-2 (G2) and Grade-3 (G3). The system uses seven layers in CNN, which includes convolution, pooling and fully connected layers. Initially, the input tile image is converted Red, Green and Blue (RGB) color channels, and then CNN approach is applied for the classification of tile images. Experimental results show the better classification accuracy of 96.17% for surface grading of ceramic tiles using a deep learning approach.

Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm.

PurposeThe aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT).MethodsCombining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python.ResultsThe periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%–91.2%) for premolars and 73.4% (95% CI, 59.9%–84.0%) for molars.ConclusionsWe demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments

For single-cell experiments, it is important to accurately count the number of viable cells in a nanoliter well. We used a deep learning-based convolutional neural network (CNN) on a large amount of digital data obtained as microscopic images. The training set consisted of 103 019 samples, each representing a microscopic grayscale image. After extensive training, the CNN was able to classify the samples into four categories, i.e., 0, 1, 2, and more than 2 cells per well, with an accuracy of 98.3% when compared to determination by two trained technicians. By analyzing the samples for which judgments were discordant, we found that the judgment by technicians was relatively correct although cell counting was often difficult by the images of discordant samples. Based on the results, the system was further enhanced by introducing a new algorithm in which the highest outputs from CNN were used, increasing the accuracy to higher than 99%. Our system was able to classify the data even from wells with a different shape. No other tested machine learning algorithm showed a performance higher than that of our system. The presented CNN system is expected to be useful for various single-cell experiments, and for high-throughput and high-content screening.

Deep learning framework for recognition of cattle using muzzle point image pattern

Animal biometrics is a frontier area of computer vision, pattern recognition and cognitive science to plays the vital role for the registration, unique identification, and verification of livestock (cattle). The existing handcrafted texture feature extraction and appearance based feature representation techniques are unable to perform the animal recognition in the unconstrained environment. Recently deep learning approaches have achieved more attention for recognition of species or individual animal using visual features. In this research, we propose the deep learning based approach for identification of individual cattle based on their primary muzzle point (nose pattern) image pattern characteristics to addressing the problem of missed or swapped animals and false insurance claims. The major contributions of the work as follows: (1) preparation of muzzle point image database, which are not publically available, (2) extraction of the salient set of texture features and representation of muzzle point image of cattle using the deep learning based convolution neural network, deep belief neural network proposed approaches. The stacked denoising auto-encoder technique is applied to encode the extracted feature of muzzle point images and (3) experimental results and analysis of proposed approach. Extensive experimental results illustrate that the proposed deep learning approach outperforms state-of-the-art methods for recognition of cattle on muzzle point image database. The efficacy of the proposed deep learning approach is computed under different identification settings. With multiple test galleries, rank-1 identification accuracy of 98.99% is achieved.