Number Of Convolutional Layers Research Articles

Dual-channel deep graph convolutional neural networks.

The dual-channel graph convolutional neural networks based on hybrid features jointly model the different features of networks, so that the features can learn each other and improve the performance of various subsequent machine learning tasks. However, current dual-channel graph convolutional neural networks are limited by the number of convolution layers, which hinders the performance improvement of the models. Graph convolutional neural networks superimpose multi-layer graph convolution operations, which would occur in smoothing phenomena, resulting in performance decreasing as the increasing number of graph convolutional layers. Inspired by the success of residual connections on convolutional neural networks, this paper applies residual connections to dual-channel graph convolutional neural networks, and increases the depth of dual-channel graph convolutional neural networks. Thus, a dual-channel deep graph convolutional neural network (D2GCN) is proposed, which can effectively avoid over-smoothing and improve model performance. D2GCN is verified on CiteSeer, DBLP, and SDBLP datasets, the results show that D2GCN performs better than the comparison algorithms used in node classification tasks.

Robust steganographic framework for securing sensitive healthcare data of telemedicine using convolutional neural network

AbstractData is the key element that runs the modern society. Large amounts of data are being released day by day as a result of many activities. The digital data is transferred through the Internet which may be vulnerable to attacks while transmitting. Especially, the medical data is observed to be of at most importance. Thus, the security of the data should be preserved while communicating through the Internet. Many techniques have been evolved for securing the data, such as Encryption which makes the data not understandable by others and techniques such as Steganography, which hides the data inside other file or document. Traditional methods of Steganography‐like LSB substitution are subjected to risks of manipulation of carrier image and poor embedding capabilities. In light of this, the authors introduce a Convolutional Neural Network‐based image steganography method that operates on DICOM images which conceals the sensitive medical data in the carrier file, thereby increasing the privacy of the sensitive data. The data considered here is brain MRI scan images. The Prep Net, Hiding Net and Reveal Net networks, each with distinct number of convolution layers are used to build the model. The sensitive data is firstly segmented to detect the Region of Interest (ROI) and then the combination of sensitive data and the carrier image are fed to the Prep Net to manipulate sensitive images in such a form that it can easily fit inside the carrier image. The Hiding Net takes the result of Prep Net and embeds the sensitive image inside the carrier image to form a stego image. The Reveal Net takes the result of Hiding Net image and then extracts the sensitive image with minimal loss. The dataset contains medical images of DICOM format taken from open source platform Kaggle that contains 3500 images of 2 classes. The results obtained show that the proposed model achieves 20 percent better output compared to traditional LSB methods. As a result, the potential for hiding the data will be expanded, as well as its security.

TrioConvTomatoNet: A robust CNN architecture for fast and accurate tomato leaf disease classification for real time application

Tomato leaf diseases are a common threat to long-term tomato protection that affects many farmers worldwide. Computer-assisted technology is more prevalent for early and accurate diagnosis of tomato leaf diseases, which may reduce the likelihood that the plant will suffer further harm. The advancement of deep learning techniques has a lot of potential for accurate classification of leaf diseases through an automated feature extraction topology. By changing the network layers and their parameters in the feature extraction topology, it is possible to extract the exact features and improve the classification accuracy. In this study, TrioConvTomatoNet, novel deep convolutional neural network architecture, is proposed, which includes a 3-series convolution layer under each stage of the feature extraction topology to capture and integrate the information from the leaf images efficiently. To improve the network's learning ability and perform accurate classifications, the proposed method incorporates the stochastic gradient descent optimizer. In order to facilitate proper learning, a separate dataset preparation strategy was followed for tomato leaf image dataset collection, which includes existing and real-time images. This research is carried out with varying the number of convolution layers to express the effectiveness of the proposed method in capturing and integrating information from the leaf images. Also, the extensive experiments on an unseen dataset collection with and without augmentation express the superiority and robustness of TrioConvTomatoNet in terms of both accuracy and speed when compared to state-of-the-art methods. The proposed method achieves remarkable accuracy of about 99.39% in disease classification while processing both existing database and real-time images, making it suitable for practical deployment in agricultural settings.

Pore characterization was achieved based on the improved U-net deep learning network model and scanning electron microscope images

The SEM image method is commonly used in the qualitative characterization of shale pores. The development of shale micro-reservoir pores can be visually observed through SEM images, but the efficiency of manual image processing is low and subjective. The introduction of deep learning greatly improves the efficiency of pore analysis. In this paper, the argon ion polishing SEM image of Longmaxi Formation shale in southern Sichuan is taken as an example. Intelligent identification and quantitative characterization of pores in shale SEM images are realized by Pore-net network model. Pore-net is based on the U-net network model. The way the model reads the data is changed so that the model does not focus on the region of interest. The number of convolutional layers of the model is increased. The Canny edge extraction algorithm is added. It not only reduces the workload of data set production, but also enhances the ability of network model to identify pores. The results show that the deep learning semantic image segmentation method is suitable for pore recognition of shale SEM images. The full convolutional neural network model is used to train the manually labeled shale SEM images, which can realize the intelligent recognition and quantitative characterization of shale SEM images. Compared with FCN and DeepLab V3+ network model, Pore-net performs better. Only 170 data sets are used to train the model. The Pore-net network model still has a good recognition effect on pores, which solves the problem of low accuracy of traditional pore recognition methods. The deviation between the porosity calculated by the Pore-net network model and the experimental data is between 12% and 19%. Compared with the porosity results calculated by the binarization method and other network models, the results calculated by Pore-net are closer to the real values, which proves that the porosity calculated by the Pore-net network model is reliable.

Comparison and analysis of the number of convolutional layers in CNN-based breast cancer detection

Breast cancer is becoming more and more common, and the mortality rate is increasing, which is indeed a serious problem. To alleviate this problem, this study applies a convolutional neural network (CNN) model to analyze breast cancer images formed by mammography. The CNN model has the property of automatically learning hierarchical feature representations from the original image data. The structure of the CNN enables it to capture spatially hierarchical features at multiple scales, including edges, textures, and objects. Specifically, this study explores the performance of CNN models on prediction tasks by varying the number of convolutional layers. This study is conducted on the DDSM mammography dataset. The experimental results demonstrate that CNN models are effective for detecting breast cancer, both in terms of accuracy and loss rate, and that more convolutional layers improve performance. In particular, the model saturates at four convolutional layers to reach the highest performance. Thus, this study helps to accelerate the efficiency of breast cancer detection and paves the way for more efficient methods in the future.

Machine Learning based Sorting of Somatic Embryos for In-Line processing in Automated SE Fluidics System

Somatic embryogenesis (SE) can be a viable method for the clonal propagation of many economically significant forest trees, particularly coniferous trees like pines and spruces. However, large-scale production of SE plants requires automation to reduce manual labor and attain cost-efficiency. The most labor-intensive step of the SE process for SE plant production is selecting and harvesting mature embryos. Embryo maturation is not a synchronized process; selecting the most developed embryos capable of continuous development is necessary. However, there needs to be more research conducted on mapping morphological features to germination-competent mature somatic embryos. This paper lays down the preliminary work of employing machine learning techniques for classifying large volumes of images of mature somatic embryos processed using an automated SE processing system based on fluidics processing referred to as SE Fluidics system. The results show that machine learning could be an alternative classification methodology instead of the traditional manual morphology-based classification process based on image analysis. The paper discusses two popular image classification techniques, namely Convolution Neural Network (CNN) and Support Vector Machine (SVM), applying them to both binary (black and white) and grayscale images. It is observed that grayscale images provide better accuracy with the SVM technique and outperform morphology-based classification in terms of processing speed (17.6% faster) across the test envelope. On the other hand, CNN-based classification shows better processing speeds only at a lower number of convolution layers. Hence, the data scientist can optimally select the number of convolution layers to get the desired accuracy-processing speed combination.

RCF-ST: RICHER CONVOLUTIONAL FEATURES NETWORK WITH STRUCTURAL TUNING FOR THE EDGE DETECTION ON NATURAL IMAGES

Context. The problem of automating of the edge detection on natural images in intelligent systems is considered. The subject of the research is the deep learning convolutional neural networks for edge detection on natural images.
 Objective. The objective of the research is to improve the edge detection performance of natural images by structural tuning the richer convolutional features network architecture.
 Method. In general, the edge detection performance is influenced by a neural network architecture. To automate the design of the network structure in the paper a structural tuning of a neural network is applied. Computational costs of a structural tuning are incomparably less compared with neural architecture search, but a higher qualification of the researcher is required, and the resulting solution will be suboptimal. In this research it is successively applied first a destructive approach and then a constructive approach to structural tuning of the based architecture of the RCF neural network. The constructive approach starts with a simple architecture network. Hidden layers, nodes, and connections are added to expand the network. The destructive approach starts with a complex architecture network. Hidden layers, nodes, and connections are then deleted to contract the network. The structural tuning of the richer convolutional features network includes: (1) reducing the number of convolutional layers; (2) reducing the number of convolutions in convolutional layers; (3) removing at each stage the sigmoid activation function with subsequent calculation of the loss function; (4) addition of the batch normalization layers after convolutional layers; (5) including the ReLU activation functions after the added batch normalization layers. The obtained neural network is named RCF-ST. The initial color images were scaled to the specified size and then inputted in the neural network. The advisability of each of the proposed stages of network structural tuning was reseached by estimating the edge detection performance using the confusion matrix elements and Figure of Merit. The advisability of a structural tuning of the neural network as a whole was estimated by comparing it with methods known from the literature using the Optimal Dataset Scale and Optimal Image Scale.
 Results. The proposed convolutional neural network has been implemented in software and researched for solving the problem of edge detection on natural images. The structural tuning technique may be used for informed design of the neural network architectures for other artificial intelligence problems.
 Conclusions. The obtained RCF-ST network allows to improve the performance of edge detection on natural images. RCF-ST network is characterized by a significantly fewer parameters compared to the RCF network, which makes it possible to reduce the resource consumption of the network. Besides, RCF-ST network ensures the enhancing of the robustness of edge detection on texture background.

Effects of the Hyperparameters on CNNs for MDD Classification Using Resting-State EEG

To monitor patients with depression, objective diagnostic tools that apply biosignals and exhibit high repeatability and efficiency should be developed. Although different models can help automatically learn discriminative features, inappropriate adoption of input forms and network structures may cause performance degradation. Accordingly, the aim of this study was to systematically evaluate the effects of convolutional neural network (CNN) architectures when using two common electroencephalography (EEG) inputs on the classification of major depressive disorder (MDD). EEG data for 21 patients with MDD and 21 healthy controls were obtained from an open-source database. Five hyperparameters (i.e., number of convolutional layers, filter size, pooling type, hidden size, and batch size) were then evaluated. Finally, Grad-CAM and saliency map were applied to visualize the trained models. When raw EEG signals were employed, optimal performance and efficiency were achieved as more convolutional layers and max pooling were used. Furthermore, when mixed features were employed, a larger hidden layer and smaller batch size were optimal. Compared with other complex networks, this configuration involves a relatively small number of layers and less training time but a relatively high accuracy. Thus, high accuracy (>99%) can be achieved in MDD classification by using an appropriate combination in a simple model.

Rekognisi Huruf Tulisan Tangan Menggunakan Convolutional Neural Network

The development of technology in the field of computer vision in recent years with the application of Technological developments in the field of computer vision in recent years with the application of convolutional neural networks have shown sophisticated performance with a high level of accuracy, such as object detection. The problem in the world of computer vision that has been looking for a solution for a long time is object classification in general images. How to duplicate the human ability to understand images, so that computers can recognize objects in images like humans. Therefore, the need for deep learning is one branch of machine learning where the algorithm used is inspired by the workings of the human brain. Some people may be more familiar with Convolution Neural Network. CNN is used to recognize and classify patterns in handwriting. The network assumes that the input used is an image. The network has a special layer called the convolution layer. In this layer, the images are inserted according to the predefined filters. In this study, various combinations of CNN architectural designs were carried out such as the number of convolution layers, stride size, number of epochs, type of kernel size optimizer. The research data comes from the National Institute of Standards and Technology (NIST) database, then the data is divided into three, namely 60% training data, 20% validation and 20% testing. The results of this experiment produce a very good accuracy value using 2 convolution layers, 50 epochs, with Adam optimizer producing an accuracy value of 99.5% when testing the model. Then evaluate the model using the confusion matrix, assigning a high value with an average value of 100% accuracy, while for the average value of precision with a value of 100%, for an average recall value of 100%, and finally an average value of f1 score of 100%.

A new real-time image semantic segmentation framework based on a lightweight deep convolutional encoder-decoder architecture for robotic environment sensing

Background: Image semantic segmentation can be understood as the allocation of a predefined category label to each pixel in the image to achieve the region segmentation of the image. Different categories in the image are identified with different colors. While achieving pixel classification, the position information of pixel points of different categories in the image is retained. Purpose: Due to the influence of background and complex environment, the traditional semantic segmentation methods have low accuracy. To alleviate the above problems, this paper proposes a new real-time image semantic segmentation framework based on a lightweight deep convolutional encoder-decoder architecture for robotic environment sensing. Methodology: This new framework is divided into three stages: encoding stage, decoding stage and dimension reduction stage. In the coding stage, a cross-layer feature map fusion (CLFMF) method is proposed to improve the effect of feature extraction. In the decoding stage, a new lightweight decoder (LD) structure is designed to reduce the number of convolutional layers to speed up model training and prediction. In the dimension reduction stage, the convolution dimension reduction method (CDR) is presented to connect the encoder and decoder layer by layer to enhance the decoder effect. Results: Compared with other state-of-the-art image semantic segmentation methods, we conduct comparison experiments on datasets Cityscapes, SUN RGB-D, CamVid, KITTI. The Category iIoU combined with the proposed method is more than 70%, and the Category IoU is as high as 89.7%. Conclusion: The results reflect that the new method can achieve the better semantic segmentation effect.

Impact of Negation and AnA-Words on Overall Sentiment Value of the Text Written in the Bosnian Language

In this manuscript, we present our efforts to develop an accurate sentiment analysis model for Bosnian-language tweets which incorporated three elements: negation cues, AnA-words (referring to maximizers, boosters, approximators, relative intensifiers, diminishers, and minimizers), and sentiment-labeled words from a lexicon. We used several machine-learning techniques, including SVM, Naive Bayes, RF, and CNN, with different input parameters, such as batch size, number of convolution layers, and type of convolution layers. In addition to these techniques, BOSentiment is used to provide an initial sentiment value for each tweet, which is then used as input for CNN. Our best-performing model, which combined BOSentiment and CNN with 256 filters and a size of 4×4, with a batch size of 10, achieved an accuracy of over 92%. Our results demonstrate the effectiveness of our approach in accurately classifying the sentiment of Bosnian tweets using machine-learning techniques, lexicons, and pre-trained models. This study makes a significant contribution to the field of sentiment analysis for under-researched languages such as Bosnian, and our approach could be extended to other languages and social media platforms to gain insight into public opinion.

ScASGC: An adaptive simplified graph convolution model for clustering single-cell RNA-seq data

Single-cell RNA sequencing (scRNA-seq) is now a successful technique for identifying cellular heterogeneity, revealing novel cell subpopulations, and forecasting developmental trajectories. A crucial component of the processing of scRNA-seq data is the precise identification of cell subpopulations. Although many unsupervised clustering methods have been developed to cluster cell subpopulations, the performance of these methods is vulnerable to dropouts and high dimensionality. In addition, most existing methods are time-consuming and fail to adequately account for potential associations between cells. In the manuscript, we present an unsupervised clustering method based on an adaptive simplified graph convolution model called scASGC. The proposed method builds plausible cell graphs, aggregates neighbor information using a simplified graph convolution model, and adaptively determines the most optimal number of convolution layers for various graphs. Experiments on 12 public datasets show that scASGC outperforms both classical and state-of-the-art clustering methods. In addition, in a study of mouse intestinal muscle containing 15,983 cells, we identified distinct marker genes based on the clustering results of scASGC. The source code of scASGC is available at https://github.com/ZzzOctopus/scASGC.

Real-time scene classification of unmanned aerial vehicles remote sensing image based on Modified GhostNet.

Unmanned Aerial Vehicles (UAVs) play an important role in remote sensing image classification because they are capable of autonomously monitoring specific areas and analyzing images. The embedded platform and deep learning are used to classify UAV images in real-time. However, given the limited memory and computational resources, deploying deep learning networks on embedded devices and real-time analysis of ground scenes still has challenges in actual applications. To balance computational cost and classification accuracy, a novel lightweight network based on the original GhostNet is presented. The computational cost of this network is reduced by changing the number of convolutional layers. Meanwhile, the fully connected layer at the end is replaced with the fully convolutional layer. To evaluate the performance of the Modified GhostNet in remote sensing scene classification, experiments are performed on three public datasets: UCMerced, AID, and NWPU-RESISC. Compared with the basic GhostNet, the Floating Point Operations (FLOPs) are reduced from 7.85 MFLOPs to 2.58 MFLOPs, the memory is reduced from 16.40 MB to 5.70 MB, and the predicted time is improved by 18.86%. Our modified GhostNet also increases the average accuracy (Acc) (4.70% in AID experiments, 3.39% in UCMerced experiments). These results indicate that our Modified GhostNet can improve the performance of lightweight networks for scene classification and effectively enable real-time monitoring of ground scenes.

CNN-based deep learning method for predicting the disease response to the Neoadjuvant Chemotherapy (NAC) treatment in breast cancer

ObjectiveThe objective of the study is to evaluate the performance of CNN-based proposed models for predicting patients' response to NAC treatment and the disease development process in the pathological area. The study aims to determine the main criteria that affect the model's success during training, such as the number of convolutional layers, dataset quality and depended variable. MethodThe study uses pathological data frequently used in the healthcare industry to evaluate the proposed CNN-based models. The researchers analyze the classification performances of the models and evaluate their success during training. ResultsThe study shows that using deep learning methods, particularly CNN models, can offer strong feature representation and lead to accurate predictions of patients' response to NAC treatment and the disease development process in the pathological area. A model that predicts ‘miller coefficient’, ‘tumor lymph node value’, ‘complete response in both tumor and axilla’ values with high accuracy, which is considered to be effective in achieving complete response to treatment, has been created. Estimation performance metrics have been obtained as 87%, 77% and 91%, respectively. ConclusionThe study concludes that interpreting pathological test results with deep learning methods is an effective way of determining the correct diagnosis and treatment method, as well as the prognosis follow-up of the patient. It provides clinicians with a solution to a large extent, particularly in the case of large, heterogeneous datasets that can be challenging to manage with traditional methods. The study suggests that using machine learning and deep learning methods can significantly improve the performance of interpreting and managing healthcare data.

Improved YOLOv7 Network Model for Gangue Selection Robot for Gangue and Foreign Matter Detection in Coal.

Coal production often involves a substantial presence of gangue and foreign matter, which not only impacts the thermal properties of coal and but also leads to damage to transportation equipment. Selection robots for gangue removal have garnered attention in research. However, existing methods suffer from limitations, including slow selection speed and low recognition accuracy. To address these issues, this study proposes an improved method for detecting gangue and foreign matter in coal, utilizing a gangue selection robot with an enhanced YOLOv7 network model. The proposed approach entails the collection of coal, gangue, and foreign matter images using an industrial camera, which are then utilized to create an image dataset. The method involves reducing the number of convolution layers of the backbone, adding a small size detection layer to the head to enhance the small target detection, introducing a contextual transformer networks (COTN) module, employing a distance intersection over union (DIoU) loss border regression loss function to calculate the overlap between predicted and real frames, and incorporating a dual path attention mechanism. These enhancements culminate in the development of a novel YOLOv71 + COTN network model. Subsequently, the YOLOv71 + COTN network model was trained and evaluated using the prepared dataset. Experimental results demonstrated the superior performance of the proposed method compared to the original YOLOv7 network model. Specifically, the method exhibits a 3.97% increase in precision, a 4.4% increase in recall, and a 4.5% increase in mAP0.5. Additionally, the method reduced GPU memory consumption during runtime, enabling fast and accurate detection of gangue and foreign matter.

A Soybean Classification Method Based on Data Balance and Deep Learning

Soybean is a type of food crop with economic benefits. Whether they are damaged or not directly affects the survival and nutritional value of soybean plants. In machine learning, unbalanced data represent a major factor affecting machine learning efficiency, and unbalanced data refer to a category in which the number of samples in one category is much larger than that in the other, which biases the classification results towards a category with a large number of samples and thus affects the classification accuracy. Therefore, the effectiveness of the data-balancing method based on a convolutional neural network is investigated in this paper, and two balancing methods are used to expand the data set using the over-sampling method and using the loss function with assignable class weights. At the same time, to verify the effectiveness of the data-balancing method, four networks are introduced for control experiments. The experimental results show that the new loss function can effectively improve the classification accuracy and learning ability, and the classification accuracy of the DenseNet network can reach 98.48%, but the classification accuracy will be greatly reduced by using the data-augmentation method. With the binary classification method and the use of data-augmentation data sets, the excessive number of convolution layers will lead to a reduction in the classification accuracy and a small number of convolution layers can be used for classification purposes. It is verified that a neural network using a small convolution layer can improve the classification accuracy by 1.52% using the data-augmentation data-balancing method.

A space and time efficient convolutional neural network for age group estimation from facial images.

Age estimation has a wide range of applications, including security and surveillance, human-computer interaction, and biometrics. Facial aging is a stochastic process affected by various factors, such as lifestyle, habits, genetics, and the environment. Extracting age-related facial features to predict ages or age groups is a challenging problem that has attracted the attention of researchers in recent years. Various methods have been developed to solve the problem, including classification, regression-based methods, and soft computing approaches. Among these, the most successful results have been obtained by using neural network based artificial intelligence (AI) techniques such as convolutional neural networks (CNN). In particular, deep learning approaches have achieved improved accuracies by automatically extracting features from images of the human face. However, more improvements are still needed to achieve faster and more accurate results. To address the aforementioned issues, this article proposes a space and time-efficient CNN method to extract distinct facial features from face images and classify them according to age group. The performance loss associated with using a small number of parameters to extract high-level features is compensated for by including a sufficient number of convolution layers. Additionally, we design and test suitable CNN structures that can handle smaller image sizes to assess the impact of size reduction on performance. To validate the proposed CNN method, we conducted experiments on the UTKFace and Facial-age datasets. The results demonstrated that the model outperformed recent studies in terms of classification accuracy and achieved an overall weighted F1-score of 87.84% for age-group classification problem.

A Novel Approach for Optimization of Convolution Neural Network with Particle Swarm Optimization and Genetic Algorithm for Face Recognition

Convolutional neural networks are contemporary deep learning models that are employed for many various applications. In general, the filter size, number of filters, number of convolutional layers, number of fully connected layers, activation function and learning rate are some of the hyperparameters that significantly determine how well a CNN performs.. Generally, these hyperparameters are selected manually and varied for each CNN model depending on the application and dataset. During optimization, CNN could get stuck in local minima. To overcome this, metaheuristic algorithms are used for optimization. In this work, the CNN structure is first constructed with randomly chosen hyperparameters and these parameters are optimized using Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) algorithm. A CNN with optimized hyperparameters is used for face recognition. CNNs optimized with these algorithms use RMSprop optimizer instead of stochastic gradient descent. This RMSprop optimizer helps the CNN reach global minimum quickly. It has been observed that optimizing with GA and PSO improves the performance of CNNs. It also reduces the time it takes for the CNN to reach the global minimum.

English

Diabetic Retinopathy (DR) is a potential condition of Diabetes Mellitus (DM) that causes lesions on the retina, reducing vision and perhaps leading to disability if not properly treated. A Special Generative Adversarial Network with Ensemble Classification Regression (SGAN-ECR) model was introduced for the effective categorisation of various DR grades with the generation of high-contrast and low-saturated RF images. However, the Multi-Scale Attention Residual Network with Gradient Boosting (MSA-ResNet-GB) model used in SGAN-ECR requires manually assigning values for many hyperparameters. An inappropriate value of the hyper-parameter leads to an increased error rate. This paper presents a method using the Enhanced Mine Blast Algorithm (EMBA) for selecting optimal hyper-parameters of MSA-ResNet-GB, such as the number of convolution layers, number of filters, filter size, number of Fully Connected (FC) layers and the hidden units in the FC layer to improve the architecture. Essential principles of EMBA are derived from the mine bomb explosion in real-time applications. The initial population of shrapnel fragments represents the initial hyper-parameter, and the computation of their subsequent locations represents the search for the best hyper-parameter. SGAN-ECR with parameter-optimised MSA-ResNet is named SGAN-OECR. In SGAN-OECR, optimised MSA-ResNet is trained by training images, and then the trained model is used to recognise the abnormalities in test images. MLP classifiers in the last FC layers of MSA-ResNet classify the severity of DR lesions. Finally, the experimental results of SGAN-OECR on Kaggle-APTOS and IDRiD datasets have 99.03% and 98.63% of accuracy, respectively, which is higher than the existing DR classification models.

Multi scale pixel attention and feature extraction based neural network for image denoising

In this paper, we propose a blind Gaussian denoising network that utilize the features of the input image and its negative for generating denoised output of the same. The proposed network is a dual path model which employs a multi-scale pixel attention (MSPA) block on one path and a multi-scale feature extraction (MSFE) block on another. The concept of using the features of a negative image (that it highlights the low contrast region) in blind Gaussian denoising network is, to the best of our knowledge, a first such attempt. The proposed MSPA and MSFE blocks are designed to focus on the features of the image at multiple scales. The MSPA block focuses on the important features of the negative of the input image whereas the MSFE block focuses on extracting features of the input noisy image. The features of both the images are then combined and a final residual noise is obtained, subtracting which from the input noisy image produces the final denoised result. The proposed network is lightweight and fast, due to the low number of convolutional layers involved, and produces superior results (both quantitatively and qualitatively) when compared with various traditional and learning based blind Gaussian denoising techniques. The code of this paper can be downloaded from https://github.com/RTSIR/NIFBGDNet.