Automated white corpuscles nucleus segmentation using deep neural network from microscopic blood smear

Objective: The antinuclear antibodies (ANA) that present in the human serum have a link with various autoimmune diseases. Human Epithelial type-2 (HEp-2) cells acts as a substance in the Indirect Immuno fluorescence (IIF) test for diagnosing these autoimmune diseases. In recent times, the computer-aided diagnosis of autoimmune diseases by the HEp-2 cell classification has drawn more interest. Though, they often pose limitations like large intra-class and small inter-class variations. Hence, various efforts have been performed to automate the procedure of HEp-2 cell classification. To overcome these problems, this research work intends to propose a new HEp-2 classification process. Materials and Methods: This is regulated by integrating two processes, namely, segmentation and classification. Initially, the segmentation of the HEp-2 cells is carried out by deploying the morphological operations. In this paper, two morphology operations are deployed called opening and closing. Further, the classification process is exploited by proposing a modified Convolutional Neural Network (CNN). The main objective is to classify the HEp-2 cells effectively (Centromere, Golgi, Homogeneous, Nucleolar, NuMem, and Speckled) and is made by exploiting the optimization concept. This is implanted by developing a new algorithm called Distance Sorting Lion Algorithm (DSLA), which selects the optimal convolutional layer in CNN. Results: Through the performance analysis, the performance of the proposed model for test case 1 at learning percentage 60 is 3.84%, 1.79%, 6.22%, 1.69%, and 5.53% better than PSO, FF, GWO, WOA, and LA, respectively. At 80, the performance of the proposed model is 5.77%, 6.46%, 3.95%, 3.24%, and 5.55% better from PSO, FF, GWO, WOA, and LA, respectively. Hence, the performance of the proposed work is proved over other models under different measures. Conclusion: Finally, the performance is evaluated by comparing it with the other conventional algorithms in terms of accuracy, sensitivity, specificity, precision, FPR, FNR, NPV, MCC, F1-Score and FDR, and proves the efficacy of the proposed model.

Alcoholic brain injury via 8-layer customized deep convolution neural network

STEMedicine ◽

10.37175/stemedicine.v2i8.97 ◽

2021 ◽

Vol 2 (8) ◽

pp. e97

Author(s):

Ziquan Zhu ◽

Mackenzie Brown

Keyword(s):

Neural Network ◽

Brain Injury ◽

Diagnostic Errors ◽

The Body ◽

Convolution Neural Network ◽

Comparison Results ◽

Deep Convolution Neural Network ◽

Sensitivity Specificity ◽

And Behavior ◽

Fully Connected

Alcohol can act quickly in the human body and alter mood and behavior. If alcohol is consumed in excess, it will accumulate in the organs of the body, especially in the liver and brain. To a certain extent, the symptoms of alcoholism will appear. So far, the main method of diagnosis of alcoholic brain injury is through MRI images by radiologists. However, this is a very subjective diagnosis. Radiologists may be affected by external factors, such as physical discomfort, lack of rest, inattention, etc., resulting in diagnostic errors. In this paper, we proposed a novel 8-layer customized deep convolution neural network for alcoholic brain injury detection, which contains five convolution layers, five pooling layers, and three fully connected layers. There are three improvements in this paper, (i) Based on deep learning, we proposed a method for automatic diagnosis of alcoholic brain injury; (ii) We introduced Dropout to the proposed structure to improve robustness; (iii) Compared with other state-of-the-art approaches, the proposed structure is more efficient. The experimental results showed that the sensitivity, specificity, precision, accuracy, F1, MCC and FMI were 96.14±1.99, 96.20±1.47, 95.98±1.54, 96.17±1.55, 96.05±1.62, 93.34±3.11, 96.06±1.62 respectively. According to comparison results, our method performed the best. The proposed model is effective in detecting alcoholic brain injury based on MRI images.

Multilane Detection Entropy-Based Fusion Model by using Iterative Seed and Optimized Deep Convolutional Neural Network

International Journal of Recent Technology and Engineering - 2 ◽

10.35940/ijrte.b1505.0882s819 ◽

2019 ◽

Vol 8 (2S8) ◽

pp. 1967-1974

Keyword(s):

Neural Network ◽

Search Algorithm ◽

Lane Detection ◽

Learning Approaches ◽

Driver Assistance ◽

Fusion Model ◽

Driver Assistance Systems ◽

Proposed Model ◽

Segmentation Approach ◽

Deep Convolution Neural Network

In today’s world, the conditions of road is drastically improved as compared with past decade. Most of the express highways are made up of cement concrete and equipped with increased lane size. Apparently speed of the vehicle will increase. Therefore there are more chances for accidents. To avoid the accidents in recent days driver assistance systems are designed to detect the various lane. The detected information of lane path is used for controlling the vehicles and giving alerts to drivers. In this paper the entropy based fusion approach is presents for detecting multi-lanes. The Earth Worm- Crow Search Algorithm (EW-CSA) which is based on Deep Convolution Neural Network(DCNN) is utilized for consolidating the outcomes. At first, the deep learning approaches for path location is prepared using an optimization algorithm and EW-CSA, which focus on characterizing every pixel accurately and require post preparing activities to surmise path data. Correspondingly, the region based segmentation approach is utilizing for the multi-lane detection. An entropy based fusion model is used because this method preserved all the information in the image and reduces the noise effects. The performance of proposed model is analyzed in terms of accuracy, sensitivity, and specificity, providing superior results with values 0.991, 0.992, and 0.887, respectively

Super-resolution Ultrasound Imaging Scheme Based on a Symmetric Series Convolutional Neural Network

10.36227/techrxiv.15066315 ◽

2021 ◽

Author(s):

Lakpa Dorje Tamang

Keyword(s):

Neural Network ◽

Feature Extraction ◽

Convolutional Neural Network ◽

Ultrasound Image ◽

Similarity Index ◽

Super Resolution ◽

Input Image ◽

Feature Maps ◽

Proposed Model ◽

Reconstruction Performance

In this paper, we propose a symmetric series convolutional neural network (SS-CNN), which is a novel deep convolutional neural network (DCNN)-based super-resolution (SR) technique for ultrasound medical imaging. The proposed model comprises two parts: a feature extraction network (FEN) and an up-sampling layer. In the FEN, the low-resolution (LR) counterpart of the ultrasound image passes through a symmetric series of two different DCNNs. The low-level feature maps obtained from the subsequent layers of both DCNNs are concatenated in a feed forward manner, aiding in robust feature extraction to ensure high reconstruction quality. Subsequently, the final concatenated features serve as an input map to the latter 2D convolutional layers, where the textural information of the input image is connected via skip connections. The second part of the proposed model is a sub-pixel convolutional (SPC) layer, which up-samples the output of the FEN by multiplying it with a multi-dimensional kernel followed by a periodic shuffling operation to reconstruct a high-quality SR ultrasound image. We validate the performance of the SS-CNN with publicly available ultrasound image datasets. Experimental results show that the proposed model achieves an exquisite reconstruction performance of ultrasound image over the conventional methods in terms of peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM), while providing compelling SR reconstruction time.

CBIR-CNN: Content Based Image Retrieval on celebrity data using Deep Convolution Neural Network

Recent Advances in Computer Science and Communications ◽

10.2174/2666255813666200129111928 ◽

2020 ◽

Vol 13 ◽

Author(s):

Pushpendra Singh ◽

P.N. Hrisheekesha ◽

Vinai Kumar Singh

Keyword(s):

Neural Network ◽

Computer Vision ◽

Deep Learning ◽

Image Retrieval ◽

Image Data ◽

Content Based Image Retrieval ◽

Facial Image ◽

Huge Amount ◽

Proposed Model ◽

Deep Convolution Neural Network

Background: Finding region of interest in an image and content-based image analysis has been a challenging task for last two decades. With the advancement in image processing, computer vision field and huge amount of image data generation, to manage this huge amount of data Content-Based Image Retrieval System (CBIR) has attracted several researchers as a common technique to manage this huge amount of data. It is an approach of searching user interest, based on visual information present in an image. The requirement of high computation power and huge memory limits deployment of CBIR technique in real-time scenarios. Objective: In this paper an advanced deep learning model is applied for CBIR on facial image data. We design a deep convolution neural network architecture where activation of convolution layer is used for feature representation and include max pooling as feature reduction technique. Furthermore, our model uses partial feature mapping as image descriptor to incorporate the property that facial image contains repeated information. Method: Existing CBIR approaches primarily consider colour, texture and low-level features for mapping and localizing image segments. While deep learning has shown high performance in numerous fields of research, its application in CBIR is still very limited. Human face contains significant information to be used in a content driven task and applicable to various applications of computer vision and multimedia systems. In this research work, a deep learning-based model has been discussed for content-based image retrieval (CBIR). In CBIR, there are two important things 1) classification and 2) retrieval of image based on similarity. For the classification purpose a four-convolution layer model has been proposed. For the calculation of the similarity Euclidian distance measure has been used between the images. Results: Proposed model is completely unsupervised, and it is fast and accurate in comparison to other deep learning models applied for CBIR over facial dataset. The proposed method provided satisfactory results from the experiment. It outperforms other CNN-based models and other unsupervised techniques used for CBIR. The proposed method provided satisfactory results from the experiment and it outperforms other CNN-based models such as VGG16, Inception V3, ResNet50 and MobileNet. Moreover, the performance of proposed model has been compared with pre-trained models in terms of accuracy, storage space and inference time.

Self-Learning Network-based segmentation for real-time brain M.R. images through HARIS

PeerJ Computer Science ◽

10.7717/peerj-cs.654 ◽

2021 ◽

Vol 7 ◽

pp. e654

Author(s):

Parvathaneni Naga Srinivasu ◽

Valentina Emilia Balas

Keyword(s):

Learning Strategies ◽

Matthews Correlation Coefficient ◽

Similarity Index ◽

Training Model ◽

Initial Assessment ◽

Jaccard Similarity ◽

Learning Network ◽

Proposed Model ◽

Sensitivity Specificity ◽

Self Learning

In recent years in medical imaging technology, the advancement for medical diagnosis, the initial assessment of the ailment, and the abnormality have become challenging for radiologists. Magnetic resonance imaging is one such predominant technology used extensively for the initial evaluation of ailments. The primary goal is to mechanizean approach that can accurately assess the damaged region of the human brain throughan automated segmentation process that requires minimal training and can learn by itself from the previous experimental outcomes. It is computationally more efficient than other supervised learning strategies such as CNN deep learning models. As a result, the process of investigation and statistical analysis of the abnormality would be made much more comfortable and convenient. The proposed approach’s performance seems to be much better compared to its counterparts, with an accuracy of 77% with minimal training of the model. Furthermore, the performance of the proposed training model is evaluated through various performance evaluation metrics like sensitivity, specificity, the Jaccard Similarity Index, and the Matthews correlation coefficient, where the proposed model is productive with minimal training.

Plant Disease Detection using Deep Learning

International Journal of Recent Technology and Engineering - 2 ◽

10.35940/ijrte.a2139.059120 ◽

2020 ◽

Vol 9 (1) ◽

pp. 909-914

Keyword(s):

Neural Network ◽

Deep Learning ◽

Plant Disease ◽

Disease Detection ◽

Virus Identification ◽

Detection Model ◽

Agriculture Industry ◽

Whole Plant ◽

Proposed Model ◽

Testing Accuracy

[Context] Plants play an essential role in climate change, agriculture industry and a country’s economy. Thereby taking care of plants is very crucial. Just like humans, plants are effected by several disease caused by bacteria, fungi and virus. Identification of these disease timely and curing them is essential to prevent whole plant from destruction. [Objective] This paper proposes a deep learning based model named plant disease detector. The model is able to detect several diseases from plants using pictures of their leaves. [Methodology] Plant disease detection model is developed using neural network. First of all augmentation is applied on dataset to increase the sample size. Later Convolution Neural Network (CNN) is used with multiple convolution and pooling layers. PlantVillage dataset is used to train the model. After training the model, it is tested properly to validate the results. [Results] We have performed different experiments using this model. 15% of data from PlantVillage data is used for testing purpose that contains images of healthy as well as diseased plants. Proposed model has achieved 98.3% testing accuracy. [Conclusion] This study is focused on deep learning model to detect disease in plant leave. But, in future model can be integrated with drone or any other system to live detect diseases from plants and report the diseased plants location to people so that they can be cured accordingly.

Super-resolution Ultrasound Imaging Scheme Based on a Symmetric Series Convolutional Neural Network

10.36227/techrxiv.15066315.v1 ◽

2021 ◽

Author(s):

Lakpa Dorje Tamang

Keyword(s):

Neural Network ◽

Feature Extraction ◽

Convolutional Neural Network ◽

Ultrasound Image ◽

Similarity Index ◽

Super Resolution ◽

Input Image ◽

Feature Maps ◽

Proposed Model ◽

Reconstruction Performance

In this paper, we propose a symmetric series convolutional neural network (SS-CNN), which is a novel deep convolutional neural network (DCNN)-based super-resolution (SR) technique for ultrasound medical imaging. The proposed model comprises two parts: a feature extraction network (FEN) and an up-sampling layer. In the FEN, the low-resolution (LR) counterpart of the ultrasound image passes through a symmetric series of two different DCNNs. The low-level feature maps obtained from the subsequent layers of both DCNNs are concatenated in a feed forward manner, aiding in robust feature extraction to ensure high reconstruction quality. Subsequently, the final concatenated features serve as an input map to the latter 2D convolutional layers, where the textural information of the input image is connected via skip connections. The second part of the proposed model is a sub-pixel convolutional (SPC) layer, which up-samples the output of the FEN by multiplying it with a multi-dimensional kernel followed by a periodic shuffling operation to reconstruct a high-quality SR ultrasound image. We validate the performance of the SS-CNN with publicly available ultrasound image datasets. Experimental results show that the proposed model achieves an exquisite reconstruction performance of ultrasound image over the conventional methods in terms of peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM), while providing compelling SR reconstruction time.

Deep Convolution Neural Network Model for Credit-Card Fraud Detection and Alert

Journal of Artificial Intelligence and Capsule Networks - September 2019 ◽

10.36548/jaicn.2021.2.003 ◽

2021 ◽

Vol 3 (2) ◽

pp. 101-112

Author(s):

Joy Iong-Zong Chen ◽

Kong-Long Lai

Keyword(s):

Neural Network ◽

Machine Learning ◽

Deep Learning ◽

Credit Card ◽

Fraud Detection ◽

Financial Fraud ◽

Detection Accuracy ◽

Learning Models ◽

Proposed Model ◽

Deep Convolution Neural Network

With the exponential increase in the usage of the internet, numerous organisations, including the financial industry, have operationalized online services. The massive financial losses occur as a result of the global growth in financial fraud. Henceforth, devising advanced financial fraud detection systems can actively detect the risks such as illegal transactions and irregular attacks. Over the recent years, these issues are tackled to a larger extent by means of data mining and machine learning techniques. However, in terms of unknown attack pattern identification, big data analytics and speed computation, several improvements must be performed in these techniques. The Deep Convolution Neural Network (DCNN) scheme based financial fraud detection scheme using deep learning algorithm is proposed in this paper. When large volume of data is involved, the detection accuracy can be enhanced by using this technique. The existing machine learning models, auto-encoder model and other deep learning models are compared with the proposed model to evaluate the performance by using a real-time credit card fraud dataset. Over a time duration of 45 seconds, a detection accuracy of 99% has been obtained by using the proposed model as observed in the experimental results.

Alcoholism via 6-layer customized deep convolution neural network

STEMedicine ◽

10.37175/stemedicine.v2i7.93 ◽

2021 ◽

Vol 2 (7) ◽

pp. e93

Author(s):

Ziquan Zhu

Keyword(s):

Neural Network ◽

Nervous System ◽

Network Structure ◽

Human Body ◽

Convolution Neural Network ◽

Brain Images ◽

Comparison Results ◽

Deep Convolution Neural Network ◽

Neural Network Structure ◽

Sensitivity Specificity

Background: Alcoholism is caused by excessive alcohol into the human body. Alcohol primarily damages the central nervous system of the human body and causes the nervous system function disorder and inhibition. Severe addiction can lead to respiratory circulation center inhibition, paralysis and even death. So far, the diagnosis of alcoholism is done by radiologist's manual CT examination. However, the diagnosis process is time-consuming, subjective and boring for doctors. External factors, such as extreme fatigue, lack of sleep and mental concentration, can easily affect the diagnosis process.Methods: In order to solve this problem, this paper proposed a new neural network based on computer vision, which used deep convolution neural network to diagnose alcoholism automatically. A total of 216 brain images were collected. In the 6-layer customized deep convolution neural network structure, there were four convolution layers and two fully connected layers, and each convolution layer was connected with a pooling layer.Results: The results showed that the accuracy, sensitivity, specificity, precision, F1, MCC and FMI were 95.96%±1.44%, 95.96%±1.66%, 95.95%±1.67%, 95.73%±1.72%, 95.84%±1.48%, 91.92%±2.87% and 95.84%±1.48% respectively.Conclusion: It can be concluded from comparison results that the proposed neural network structure is more effective than four state-of-the-art approaches. The proposed method has high accuracy and can be used as a diagnostic method for alcoholism.