Search

Pose estimation is a fascinating aspect of pattern recognition that is used in a variety of industries, such as technology, healthcare, gaming, etc. The goal of human posture estimation is to foretell the positions of joints and body components in still photos and moving pictures. Given the wide range of uses for such a system, it constitutes one of the most intriguing domains of computer vision research and has had significant growth. The compact convolutional neural networks architecture for human posture estimation proposed in this study, called BlazePose, is intended towards mobile device legitimate prediction. During analysis, the network operates from over 30 fps on something like a Pixel 2 mobile, providing 33 body keypoints for such a specific individual. As a result, it is particularly well-suited for real-time applications like as fitness tracking and action recognition comprehension. This paper makes two major contributions: a novel body posture tracking approach and a lightweight body pose estimation computational model that evaluates parameters utilizing heatmaps and regression. The suggested method has improved estimate accuracy, which is useful when it is used for robotics, augmented reality, sports and fitness, action recognition, motion analysis, and human-computer interface. [ABSTRACT FROM AUTHOR]

The deaths caused by road accidents are one of India's most pressing issues. The inattentiveness of the driver is responsible for almost 80% of all collisions. Mobile phone use, conversing with passengers, reaching behind the wheel to grab something, and drinking while driving are just a few of the factors that can cause a driver to become distracted. This project aims to create a reliable and effective system for detecting and classifying distracted drivers. The efficacy of CNNs in the field of computer vision was the inspiration for this paper, which proposes a paradigm based on Convolutional Neural Network which not only senses but also recognizes the driver who is distracted. For this purpose it is using a publicly available dataset from the Kaggle website as an input for this model. There are several different forms of distractions, but this study focuses on the manual distraction, which is focused on the driver's stance. The goal of this work is to keep an eye on the driver's actions and seek for diversion signs. Distractions are divided into two groups in the model: safe driving and unsafe or distracted driving. For this purpose, various architectures of CNN can be used. This work, however, employs sequential architecture and achieved the highest observed accuracy of 99% with a total of 9.5M trainable parameters. The primary goal of this research is to classify a test image into two states of the driver that have been considered. [ABSTRACT FROM AUTHOR]

State-of-the-art solutions in the areas of language modelling, generating text, speech recognition, generating image descriptions or video tagging, have been using Recurrent Neural Networks as the foundation for their approaches. Understanding the underlying concepts is therefore of tremendous importance if we want to investigate frontier methods and develop new approaches for building smart digital solutions. This paper proposes a smart city control system architecture based on deep convolutional neural networks. The control system has a multilayer architecture that combines loosely coupled intelligent components. As the main layer, a solution based on deep learning technology is applied, which allows solving several tasks simultaneously: segmentation, detection and classification of images received from surveillance cameras of the smart city system. The data obtained at the output of this layer is used for further analysis and decision-making in the smart city system. The proposed architecture has a high degree of modularity and allows the replacement of individual elements in a loosely coupled architecture. In this paper, deep learning and computer vision technologies are also considered, on the basis of which the image processing layer from video cameras is implemented. A masked recurrent neural network is used for this task. [ABSTRACT FROM AUTHOR]

Cancer is one of the most dangerous diseases which is the cause of death of around half a million people every year. Research scientists and doctors are still finding methods to avoid these casualties. Brain tumor and Liver lesions are among the most fatal cancers and early detection of these diseases can help in reducing large human loss every year. Various methods are used for early detection of brain tumor and liver lesion from medical images using computer vision and image processing. This paper provides a detailed analysis on various methods used for detection and segmentation of brain tumor and liver lesion. Due to the large availability of image data and complex computing resources which can process them has made a deep learning approach popular for medical image analysis. Deep learning architectures like Convolutional Neural Network (CNN), Fully Convolutional Network, Auto Encoders and Generative Adversarial Network (GAN) has transformed the research in brain tumor and liver lesion analysis. This paper discusses these deep learning architectures in brief along with the various variations in these architectures for medical image analysis. The datasets used for brain tumor and liver lesion analysis are also listed and compared in the paper. The advantages, limitation and research challenges of using deep learning methods is discussed in detail along with the possible solutions for overcoming these limitations. This paper will explore possible research gaps in deep learning in terms of analyzing the medical image with different modalities, resource computation, ensemble of CNN and optimization of the deep learning model for classification and prediction of medical data. [ABSTRACT FROM AUTHOR]

Pothole detection is a challenging task as it involves inhomogeneity in intensity, complex topologies, contrast less, and noise rich background. Still, it's a very important task concerning the maintenance of roads. Potholes are being detected using computer vision which generally involves analysis of a two-dimensional image of the road or modeling of the road surface. Provided both these methods are independently used. The accuracy of this method of pothole detection is far from the top point. These bring the usage of digital image processing which develops the detection of potholes with at most accuracy. Digital Image processing provides strong receptivity, usage of huge actions, and broad benefits. In this paper, we propose a system that supports digital image processing techniques of high efficiency to result in consistent and accurate detection of potholes. This project studies an intelligent digital industrial pothole detection mechanism to increase human safety, detection capability and decrease the factor of risk. Firstly, the sources of images from the road are pre-processed to remove noise and other irrelevant features. Then a training model was developed to train the system to classify between path holes and normal cracks using convolution neural networks and MATLAB. Also, an IoT device is used in this system which updates the detected potholes on the road to the respective web server along with the location coordinates, as a part to notify a complaint to check. This system works on providing a better pothole detection technique and also updating the same to the concerned. [ABSTRACT FROM AUTHOR]

A bigger role in our lives is being played by deep learning (DL). It has already had a significant effect in fields including self-driving cars, speech recognition, cancer diagnosis, precision medicine, and forecasting. Generative adversarial networks, convolutional neural networks, and visual geometry groups employing computer vision are a few of the key facilitators of DL algorithms. Using the defined keywords "deep learning," "image quality," and "computer vision and their application," the current study conducts an exhaustive search of all pertinent research articles on deep learning in image enhancement in three easily accessible databases, namely IEEE, ScienceDirect, ResearchGate, and Kaggle. Two categories are used to separate the final collection of papers on deep learning's application to image improvement. In the first category, you'll find publications that made an effort to build and construct deep learning, especially those that focused on novel architecture, system designs, frameworks, schemes, models, techniques, and algorithms. The second group of studies includes those that made an effort to assess and examine the use of generative adversarial networks in picture improvement. [ABSTRACT FROM AUTHOR]

This paper discusses the importance of communication for human interaction, especially for individuals with hearing impairments who rely on sign language. However, learning sign language is not easy, and not many people are familiar with it. To bridge the communication gap between deaf individuals and the general public, this study focuses on developing a deep convolutional neural network (CNN) model to recognize sign language patterns, specifically SIBI Sign Language used for formal activities. The VGG16 architecture is employed, which has been trained on a sign language dataset with 1440 images. The results show an accuracy of over 98.32% in training data. This research demonstrates how technology, particularly deep learning and computer vision, can help solve communication problems and improve accessibility for individuals with hearing impairments. [ABSTRACT FROM AUTHOR]

Computer vision is an area of machine learning that enables computers and systems to take actions and make recommendations based on information from digital images, videos, and other visual inputs. The purpose of this paper is to discuss the selection of optimal parameters in car classification training. This paper discusses computer vision training using the CNN algorithm on vehicle miniature classification. One of the factors that affect the performance of Convolutional Neural Network (CNN) is the selection of appropriate parameters. The data we use comes from the primary data we collect. We took photos of 3 types of vehicles, namely buses, cars, and motorcycles with a total of 486 instances. Next we do data preprocessing by resizing 256*256px. Then we conducted various trials of data augmentation i.e: including "horizontal", "vertical", "horizontal and vertical". We experimented with different augmentations to get the highest accuracy. We present several variations of the epoch to identify the best model of CNN. The accuracy results obtained are 97.34%. [ABSTRACT FROM AUTHOR]

In the recent times, deep learning has gained significant importance in various domains. Computer Vision is one the domains in which deep learning has seen vast development and is still gaining a lot of attention. CNNs are deep architectures that are specifically applied for image classification and object detection in computer vision. Image classification has also been successfully applied in the field of agriculture. Using deep learning, various agricultural problems have been addressed and solved. Keeping in view the enormous uses of deep learning in agricultural domain, this paper gives a brief review of the various works proposed by the researchers working in the field of deep learning and agriculture. This paper will put some light on the various agricultural problems, the deep learning models that have been employed to solve the problem, any pre-processing techniques that have been applied on the images to prepare the dataset for the specific model and the results that have been achieved by the model in solving the problem they focus on. It will also provide any research gap found in the specific work. This work will be useful for future researchers to gain an overview of the work done so far and the deep learning techniques used in agriculture. [ABSTRACT FROM AUTHOR]

The world is filled with plants. They are a vital part of our ecosystem, providing food, shelter, and other services. Some objects have one use, while others have multiple uses. In this paper, we explore plants that have potential medical uses. Some of these plants have been used historically for medicinal purposes, while others may have new potential for treating health problems. Computer vision researchers have developed identification systems that help botanists pinpoint and unexplored plant genus more swiftly. Copious research has anchored on methods that magnify the utility of leaf databases for plant predictive modelling, but this can lead to leaf features that change depending on the data and extraction methodology used. We use Convolutional Neural Networks (CNNs) to unsheathe useful counsel from raw data representations. This allows us to understand the underlying structure of the data more easily. In this, we have used Medicinal leaf dataset from Mendeley Data which consist of 1500 data out of which 1471 data were used to train and validate the model. We were able to pull off an accuracy of 99.10 %. [ABSTRACT FROM AUTHOR]

As part of computer vision theory, study and application, object detection and facial recognition continue to be important. In recent years, Deep Learning and specifically Convolutional Neural Networks (CNN) have elicited a great deal of advancement. Also, has consequently gotten a lot of consideration on the worldwide phase of examination about PC vision. The Internet of things assigns actual articles that are inserted with sensors, regulation capacity, programming, and different advancements that associate and trade information with different gadgets and frameworks over the Internet or different foundations organizations. Applying profound brain organizations to IOT gadgets could achieve an age of requesting equipped for performing complex detecting and acknowledgment errands to help another domain of collaborations among people and their actual airs. We have involved a high-level processor Raspberry pi for viable speed and remote capacities. This paper includes the plan of elements for depicting the article and facial qualities followed by reconciliation with classifiers. Haar Cascade classifier is a successful for articles and face identification method. It is an item openness calculation used to distinguish faces in a picture or a constant video. [ABSTRACT FROM AUTHOR]

In recent times, several fields that exist in engineering have been heavily dominated through the integration of artificial intelligence, however machine vision systems are simultaneously being studied and observed, they encapsulate certain tools that help in deriving technical information such as the working, errors and relative faults. Such systems play a vital role in manufacturing production lines, where important factors such as efficiency and productivity are sought to be at optimum levels. This paper captures the essence of using convolutional neural networks and also the addition of VGG-19 that are used in the detection and inspection of the various type of bottle caps. The respective bottle cap classes that have been pre-determined include normal cap, no cap a misplaced cap. The implementation of this system should enable us to determine whether a bottle with a specific cap class should be rejected or accepted. High metrics of precision and accuracy of the system should be observed and beneficial to the existing and upcoming processes. [ABSTRACT FROM AUTHOR]

Many countries use the AFIS (Automated Fingerprint Identification System) to identify individuals via their fingerprints. In this system, fingerprint inputs are compared with all the images stored in the given database. Many comparisons are required when the database is enormous; thus, fingerprint classification is used to eliminate many comparisons (reduced database penetration rate). Deep learning techniques have extremely advanced several domains, especially computer vision processing. This paper surveys research on fingerprint classification utilizing deep convolutional learning. It provides an overview of convolutional deep neural networks and some types of CNN models. Moreover, it presents a comparison and complete analyses of 10 research projects based on: CNN models, dataset, and the accuracy achieved hyperparameters. The most common CNNs used in fingerprint classification are AlexNet with an accuracy of 99.79%, VGG-16 with an accuracy of 98.30%, and ResNet with 95.7%. [ABSTRACT FROM AUTHOR]

Artificial neural networks are successfully applied to many different problems. Among them, a large class of problems related to computer vision can be distinguished. In this area, the use of convolutional neural networks is particularly successful. Most of the existing neural network architectures are trained on large clusters that require a large amount of computational resources. Therefore, urgent is the task of optimizing neural networks, which can include both increasing performance and reducing the size of the computing power used. In this paper, we propose a method for optimizing (increasing the performance and reducing the amount of consumed resources) of a convolutional neural network, applicable in conditions of redundancy in the input data. Using the Caltech256 dataset and VGG16 network architecture, it was shown that the proposed method can improve network performance by 10% while maintaining accuracy and reducing the amount of resources consumed by 25%. [ABSTRACT FROM AUTHOR]

CNN models are widely utilized because they have shown outstanding results in various fields, including computer vision and speech recognition. This success has been aided by the widespread availability of capable underlying hardware platforms. Currently, hardware specialization may expose us to new architectural solutions that outperform general-purpose computers for the laser operations involved. Although different applications need different performance parameters, they place a high value on speed and energy efficiency. Meanwhile, processing has resurged as a result of its inherited high speed and high-power features. This paper looks at the potential of using CNNs with a suggested accelerator based on a modified filter algorithm. Our results show that pipeline accelerators may increase energy efficiency while competing with state-of-the-art electronic platforms in terms of power and speed. The proposed design is Xilinx ISE synthesizable and simulated. [ABSTRACT FROM AUTHOR]

The theory of pattern recognition is used in computers, in city's video surveillance cameras, when recognizing car license plates of the violators, for unlocking of smartphones, in drones, in security systems, etc. The paper is considering the solution to the problem of using image recognition methods for detecting road's signs. The purpose of the research is to develop a mobile application for the recognizing of road signs. The methodology includes a description of the stages of the video sequence for recognizing of road signs, choosing methods for their localization, classification and data storage in recognition systems. A significant result of the research is the development of a model based on convolutional neural networks for the classification of road signs. The developed model is used in a mobile application applying the methods of the libraries of computer vision. The application can be used by car drivers with installed file on the device and an autofocus camera. Testing of the application confirms its ability to perform the assigned tasks. In conclusion, findings are drawn about the effectiveness of the application. And options are proposed for improving the model for recognizing of road signs. [ABSTRACT FROM AUTHOR]

Timely analysis and recognition of leaf diseases in plants ensures good quality and productivity of crops in the field of agriculture. Recognition of rice leaf diseases is one of the biggest challenges for the farmers due to incomplete awareness regarding latest sophisticated techniques in the field of leaf disease identification. Much of the growth of rice is disturbed due to the prevalence of rice leaf diseases. Earlier detection of rice leaf diseases done manually by farmers which is very laborious and time consuming. However, requisite of automatic identification of rice leaf diseases helps farmers to save their agricultural crops more efficiently. Innovations in the field of computer vision and deep learning meets the expectations and proves to be the best solution for classification of rice leaf diseases using convolutional neural networks using samples of rice leaf diseases. In this review paper, the major focus is on performance analysis of detection of rice leaf diseases based on the architectures employed. [ABSTRACT FROM AUTHOR]

Artificial intelligence & Deep Learning plays a vital role in this era. It is doing great progress in all aspects. "Automatic Image Captioning" is a Deep Learning mechanism; its primary aim is to generate the caption for an input Image by the machine. This mechanism initially identify the object in the image by using computer vision (CV) and generate the concise meaning of an image using Natural Language processing (NLP). In this paper we use Convolution Neural Network (CNN), which takes an image as input and assigns importance (suitable biases and weights) to varies object in the given image and it can distinguish one object from other and Long Short Term Memory (LTSM) give the proper meaning to each object in the image and gives a proper description to an entire image. [ABSTRACT FROM AUTHOR]

Deep Learning is invading almost all sector of real world. With the rapidly growing demand and need of several industries, a Convolutional neural network (CNN) has become a major driving force in all image based applications. The advances in Computation infrastructures like cloud, GPUs and data science, solution to almost any problem is made simple as ever before. The formidable growth of these technologies aid in the high level understanding of computers resulting in automation and digitization while meeting the extremity of real world applications. In this paper we train the CNN to identify different types of basic skin diseases, source of cause and also attempt to predict the recurrence of the disease in future. Based on the appearance, the neural network is trained to understand the different categories of diseases and analyze the reason for occurrence and also predicts the recurrence of the disease in future. According to "A new Journal of the European Academy of Dermatology and Venereology "states that skin problems are the most commonly caused human illness. But most of the people give poor importance and doesn't consult the physician. Skin problems are often considered to be least important due to various societal and personal factors. While most of the people avoid consulting a physician due to lack of awareness, reluctant and seriousness, diagnosing the problem via images may become a good fortune to many people suffering from various skin problems. Image classification and image captioning, the subfields of computer vision are used for classifying the images and labeling them according to the visual content. In addition to these, Transfer Learning advancement in the field of Deep learning is used to identify the reason for recurrence of skin problems. Automated detection of skin problems, instance segmentation and the subsequent quantification (severity of skin problem) will be able to provide timely help to patients. Experimental results provide greater accuracy in the field of dermatology while detecting the disease, source of occurrence. Further attempts are made to predict the recurrence. [ABSTRACT FROM AUTHOR]

Face Mask recognition has developed as an extremely admired issues in several domains such as image processing, computer vision and artificial intelligence. Several novel approaches are being developed through deep learning based Convolutional Neural Network model to detect the face mask identification. In this paper, we propose "FMNet" model which indicates Face Mask detection Neural Network model for recognition of face mask person on publicly available face masks images as resources. Furthermore, we applied image preprocessing, feature extraction have performed using CNN, applied FMNet model for face mask recognition from images, and finally perform classification technique to categorize the images as person "wearing masks" and "not wearing masks". Our experimental outcome generates better performance in finding face mask by achieving accuracy as 99.7%. Moreover, comparing our proposed algorithm results with pre-trained models namely VGG 16 attains 99.6%. This system has capability to perform in real time application creates it applicable to identify people in airports, buses, schools etc. [ABSTRACT FROM AUTHOR]

The major advancing techniques in machine learning are mainly two, they are deep learning and computer vision. The advanced deep learning techniques are highly promising to increase the interest in research within the upcoming years. This is often because the eminent benefits in overcoming the drawbacks within the outdated techniques for producing the result accurately. The theme of this paper is to provide a comprehensive description on the convolution neural network and its recent improvements which includes the CNN – S convolution neural network segmentation, CNN – CBIR convolution neural network – content-based image retrieval system. This survey paper provides a detailed summary within the latest advancements in the domain of CNN with various extended applications through its classification for improved understanding. Analysing the performance is done considering the speed, accuracy and ease. [ABSTRACT FROM AUTHOR]