Operation In Computer Vision Research Articles

Visual Saliency Detection using Deep Learning

Salient object discovery models mimic the gesture of human beings and capture the most salient region/ object from the images or scenes. This field has numerous important operations in both computer vision and pattern recognition tasks. Despite hundreds of models proposed in this field, it still has a large room for exploration. This paper demonstrates a detailed overview of the recent progress of saliency discovery models in terms of heuristic- grounded ways and deep literacy- grounded ways. We've bandied and reviewed it’s co-related fields, similar as Eye obsession- vaticination, RGBD salient- object- discovery, co- saliency object discovery, and videotape- saliency- discovery models. Image saliency object discovery can fleetly prize useful information from image scenes and further assay it. At present, the traditional saliency target discovery technology still has the edge of outstanding target that can not be well saved. Convolutional neural network( CNN) can prize largely general deep features from the images and effectively express the essential point information of the images. This paper designs a model which applies CNN in deep saliency object discovery tasks. It can efficiently optimize the edges of focus objects and realize largely effective image saliency discovery through multilayer nonstop point birth, refinement of layered boundary, and original saliency point emulsion. The experimental result shows that the proposed system can achieve further robust saliency discovery to acclimate itself to complex background terrain.

A comprehensive end-to-end computer vision framework for restoration and recognition of low-quality engineering drawings

The digitization of engineering drawings is crucial for efficient reuse, distribution, and archiving. Existing computer vision approaches for digitizing engineering drawings typically assume the input drawings have high quality. However, in reality, engineering drawings are often blurred and distorted due to improper scanning, storage, and transmission, which may jeopardize the effectiveness of existing approaches. This paper focuses on restoring and recognizing low-quality engineering drawings, where an end-to-end framework is proposed to improve the quality of the drawings and identify the graphical symbols on them. The framework uses K-means clustering to classify different engineering drawing patches into simple and complex texture patches based on their gray level co-occurrence matrix statistics. Computer vision operations and a modified Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) model are then used to improve the quality of the two types of patches, respectively. A modified Faster Region-based Convolutional Neural Network (Faster R-CNN) model is used to recognize the quality-enhanced graphical symbols. Additionally, a multi-stage task-driven collaborative learning strategy is proposed to train the modified ESRGAN and Faster R-CNN models to improve the resolution of engineering drawings in the direction that facilitates graphical symbol recognition, rather than human visual perception. A synthetic data generation method is also proposed to construct quality-degraded samples for training the framework. Experiments on real-world electrical diagrams show that the proposed framework achieves an accuracy of 98.98% and a recall of 99.33%, demonstrating its superiority over previous approaches. Moreover, the framework is integrated into a widely-used power system software application to showcase its practicality. The reference codes and data can be found at https://github.com/Lattle-y/AI-recognition-for-lq-ed.git Future work will focus on improving the generalizability of the proposed framework to different quality degradation scenarios and extrapolating the application to different engineering domains.

Diabetic retinopathy prediction based on vision transformer and modified capsule network

Diabetic retinopathy is considered one of the most common diseases that can lead to blindness in the working age, and the chance of developing it increases as long as a person suffers from diabetes. Protecting the sight of the patient or decelerating the evolution of this disease depends on its early detection as well as identifying the exact levels of this pathology, which is done manually by ophthalmologists. This manual process is very consuming in terms of the time and experience of an expert ophthalmologist, which makes developing an automated method to aid in the diagnosis of diabetic retinopathy an essential and urgent need. In this paper, we aim to propose a new hybrid deep learning method based on a fine-tuning vision transformer and a modified capsule network for automatic diabetic retinopathy severity level prediction. The proposed approach consists of a new range of computer vision operations, including the power law transformation technique and the contrast-limiting adaptive histogram equalization technique in the preprocessing step. While the classification step builds up on a fine-tuning vision transformer, a modified capsule network, and a classification model combined with a classification model, The effectiveness of our approach was evaluated using four datasets, including the APTOS, Messidor-2, DDR, and EyePACS datasets, for the task of severity levels of diabetic retinopathy. We have attained excellent test accuracy scores on the four datasets, respectively: 88.18%, 87.78%, 80.36%, and 78.64%. Comparing our results with the state-of-the-art, we reached a better performance.

A proposed reconstruction method of a 3D animation scene based on a fuzzy long and short-term memory algorithm.

With the development of computer technology leading to a broad range of virtual technology implementations, the construction of virtual tasks has become highly demanded and has increased rapidly, especially in animation scenes. Constructing three-dimensional (3D) animation characters utilizing properties of actual characters could provide users with immersive experiences. However, a 3D face reconstruction (3DFR) utilizing a single image is a very demanding operation in computer graphics and vision. In addition, limited 3D face data sets reduce the performance improvement of the proposed approaches, causing a lack of robustness. When datasets are large, face recognition, transformation, and animation implementations are relatively practical. However, some reconstruction methods only consider the one-to-one processes without considering the correlations or differences in the input images, resulting in models lacking information related to face identity or being overly sensitive to face pose. A face model composed of a convolutional neural network (CNN) regresses 3D deformable model coefficients for 3DFR and alignment tasks. The manuscript proposes a reconstruction method for 3D animation scenes employing fuzzy LSMT-CNN (FLSMT-CNN). Multiple collected images are employed to reconstruct 3D animation characters. First, the serialized images are processed by the proposed method to extract the features of face parameters and then improve the conventional deformable face modeling (3DFDM). Afterward, the 3DFDM is utilized to reconstruct animation characters, and finally, high-precision reconstructions of 3D faces are achieved. The FLSMT-CNN has enhanced both the precision and strength of the reconstructed 3D animation characters, which provides more opportunities to be applied to other animation scenes.

A Sparse Local Binary Pattern extraction algorithm applied to event sensor data for object classification

Recently, new sensors with active pixels were brought to market. These sensors export local variations of light intensity in the form of asynchronous events with low latency. Since the data output format is a stream of addressable events and not a complete image of light intensities, new algorithms are required for known problems in the field of Computer Vision, such as segmentation, VO, SLAM, object, and scene recognition. There are some proposed methodologies for object recognition using conventional methods, convolutional neural networks, and third-generation neural networks based on spikes. However, convolutional neural networks and spike neural networks require specific hardware for processing, hard to miniaturize. Also, several traditional Computer Vision operators and feature descriptors have been neglected in the context of event sensors and could contribute to lighter methodologies in object recognition. This paper proposes an algorithm for local binary pattern extraction in sparse structures, typically found in this context. This paper also proposes two methodologies using local binary patterns to captures with event-based sensors for object recognition. The first methodology exploits the known motion performed by the sensor, while the second is motion agnostic. It is demonstrated experimentally that the LBP operator is a fast and light alternative that enables variable reduction using PCA in some cases. The experiments also show that it is possible to reduce the final feature vector for classification by up to 99,73% when compared to conventional methods considered state-of-the-art while maintaining comparable accuracy.

Enhancing Semantic Segmentation: Design and Analysis of Improved U-Net Based Deep Convolutional Neural Networks

In this research, we provide a state-of-the-art method for semantic segmentation that makes use of a modified version of the U-Net architecture, which is itself based on deep convolutional neural networks (CNNs). This research delves into the ins and outs of this cutting-edge approach to semantic segmentation in an effort to boost its precision and productivity. To perform semantic segmentation, a crucial operation in computer vision, each pixel in an image must be assigned to one of many predefined item classes. The proposed Improved U-Net architecture makes use of deep CNNs to efficiently capture complex spatial characteristics while preserving associated context. The study illustrates the efficacy of the Improved U-Net in a variety of real-world circumstances through thorough experimentation and assessment. Intricate feature extraction, down-sampling, and up-sampling are all part of the network's design in order to produce high-quality segmentation results. The study demonstrates comparative evaluations against classic U-Net and other state-of-the-art models and emphasizes the significance of hyperparameter fine-tuning. The suggested architecture shows excellent performance in terms of accuracy and generalization, demonstrating its promise for a variety of applications. Finally, the problem of semantic segmentation is addressed in a novel way. The experimental findings validate the relevance of the architecture's design decisions and demonstrate its potential to boost computer vision by enhancing segmentation precision and efficiency.

Animal Movement Detection along with Differentiation and Alert Processing in Farm Lands

Every life on land depends on food for living. The source of major availability of food is agriculture. Farmers and agricultural lands are exposed to lot more difficulties currently. One among them is intruders in farm fields. Animal intruders cause a tremendous amount of destruction to crops on field. The proposed system helps in detecting large animal movements using frame differentiation in OpenCV. Motion detection algorithm helps in detecting the movements of animals. OpenCV deals with computer vision operations. Frame differentiation is used to fetch differences between still and movable frames. Working on with Movidius Neural compute stick connected to raspberry Pi which used to fetch the category of animal approaching using image classifier and corresponding actions could be taken on time. Once motion is detected and alert sound is generated and hence the owners could be made aware of happenings. Movidius stick is used to learn artificial intelligence at the edge. Other applications involve, motion detection in military boundaries.

Network anomaly detection using deep learning techniques

Convolutional neural networks (CNNs) are the specific architecture of feed-forward artificial neural networks. It is the de-facto standard for various operations in machine learning and computer vision. To transform this performance towards the task of network anomaly detection in cyber-security, this study proposes a model using one-dimensional CNN architecture. The authors' approach divides network traffic data into transmission control protocol (TCP), user datagram protocol (UDP), and OTHER protocol categories in the first phase, then each category is treated independently. Before training the model, feature selection is performed using the Chi-square technique, and then, over-sampling is conducted using the synthetic minority over-sampling technique to tackle a class imbalance problem. The authors' method yields the weighted average f-score 0.85, 0.97, 0.86, and 0.78 for TCP, UDP, OTHER, and ALL categories, respectively. The model is tested on the UNSW-NB15 dataset.

875 Using Stress Testing to Identify Vulnerabilities in Artificial Intelligence Models for the Identification of Culprit Carotid Lesions in Cerebrovascular Events

Abstract Introduction Carotid atherosclerosis is a major risk factor for ischaemic stroke, a leading cause of death. Carotid CT angiography (CTA) is commonly performed following a stroke or transient ischaemic attack (TIA) to help guide patient management in secondary prevention of stroke. Deep learning algorithms can help extract greater information from scans. Method The dataset comprised CTA scans from 40 culprit and 40 non-culprit carotid arteries of patients with recent stroke/TIA, and 40 carotid arteries of asymptomatic patients without previous stroke/TIA. A 3D convolutional neural network was trained to classify carotid artery type. Each input comprised 14 axial CTA carotid patches (centred around the carotid artery) concatenated together to form a 3D volume (capturing ∼3cm of artery). 75% of the dataset was used for training and 25% for internal validation. Following training, computer vision operations were applied to input images to assess their impact on the model’s classification decisions. Results The model achieved 100% accuracy on the training set and 67% on the internal validation set. However, after subjecting input images to image operations, vulnerabilities in the deep learning model were revealed, even when using input images from the training set. For example, using a Gaussian blur filter with sigma 1.0 was sufficient to change classification decisions, as was horizontally flipping the image. Conclusions Deep learning has exceptional capabilities for learning, however the risk with such high-capacity models is failure to learn relevant features from the data. Stress testing provides a viable method to further evaluate deep learning models before clinical deployment.

A novel interdependence based multilevel thresholding technique using adaptive equilibrium optimizer

The multilevel thresholding is one important operation in computer vision, which is a subfield of artificial intelligence (AI), used to understand and interpret the data in the real world. The existing entropic methods, based on the histogram of an image for the multilevel thresholding mostly deal with the maximization of the entropic information excluding the shred boundary, which reduces the accuracy. These problems lead to the poor thresholding accuracy and a lower speed. To address the problem, we propose a novel interdependence based technique that uses the shred boundary, which is a minimization problem. A firsthand objective function is investigated, which takes care of the shred boundary. The traditional multilevel thresholding techniques are computationally expensive due to the exhaustive search process, an alternate method is to use the evolutionary computation based on a nature-inspired algorithm. In this paper, a new optimizer called adaptive equilibrium optimizer (AEO) is also proposed for multilevel thresholding, an improvement over the basic equilibrium optimizer (EO) by implementing an adaptive decision making of dispersal for nonperformer search agents. The AEO performance is compared with state-of-the-art algorithms — equilibrium optimizer (EO), gray wolf optimizer (GWO), whale optimization algorithm (WOA), squirrel search algorithm (SSA) and the wind driven optimization (WDO) algorithm, using standard benchmark functions. Based on the qualitative and quantitative analysis, the AEO outperformed EO, GWO, WOA, SSA, and WDO. The optimal thresholds are obtained by minimizing the objective function using the AEO. For the experiment, 500 images of the BSDS 500 dataset are considered. Popular metrics such as the peak signal to noise ratio (PSNR), structural similarity index (SSIM), and the feature similarity index (FSIM) are considered for a quantitative analysis. Remarkable differences in the thresholding accuracy are observed with a simultaneously decreasing computational complexity. The merits of the paper are highlighted to ensure its future use in the world of engineering applications using soft computing, a subfield of the AI.

Machine Vision Techniques for Improve Rice Grain Quality Analyzing Process

Rice grain quality inspection is a major process in rice production. To provide quality and accurate results in rice grain analyzing it is important to analyze rice grains one by one in a testing sample. In the current situation, most of rice grain producers inspect rice grains manually without using any automated process. The major problem is the accuracy of testing results depends on human quality because manually processes include human errors. The manual inspection of rice grains is a very complicated and time-consuming process due to these reasons most of the inspector's effect by external factors such as fatigue, tension etc. In this research, we provide a time-efficient and low-cost solution for reducing above-mentioned limitations by developing software. It uses modern image processing to analyze rice grains one by one efficiently over the manual examination. The quality of rice samples can be determined with the help of colour, and geometric features such as area, maximum length, maximum width and aspect ratio. This analyzing system designed and developed for measure area, maximum length, maximum width and aspect ratio by using Java programming language, morphological and colour operations in computer vision and finally the accuracy of the system tested by comparing manually tested sample and results from the system. According to the results, it shows this system provides more than 85 percent accuracy with confirming this was a better solution

Modified artificial bee colony algorithms for solving multiple circle detection problem

Determining circular shapes from digital images is one of the most required operations in computer vision and its applications. Various techniques including evolutionary and swarm intelligence-based algorithms have been introduced and successfully used over the past decade for detecting single or multiple circles. In this study, two different circle candidate list generation approaches that are based on generating a combination between abandoned food sources and the obtained food sources in the final colony have been proposed and integrated into the workflow of the standard implementation of the artificial bee colony (ABC) algorithm. Experimental studies on a set of real and synthetic images showed that the proposed list generation approaches improved the solving capabilities of ABC algorithm and decreased the total error values related to the discovered circles compared to the ABC-based implementation for which a circle or circles are selected from a candidate list containing only abandoned food sources, genetic algorithm, bacterial foraging optimization algorithm and an improved version of the Hough transform-based circle detection technique called randomized Hough transform.

STUDYING 'LIVE' CROSS-PLATFORM CIRCULATION OF IMAGES WITH A COMPUTER VISION API: AN EXPERIMENT BASED ON A SPORTS MEDIA EVENTd'Andréa, Carlos Frederico de Brito (Federal University of Minas Gerais, Brazil) <carlosfbd@gmail.com>

Considering the importance of cross-platform circulation of web contents for digital methods-oriented research, in this study we aim to expand the types of digital objects used as ‘traffic tags’ by focusing on static images as traces of online associations. We pursue this goal through a methodological experiment with Google Cloud Vision API, a computing framework for visual content analysis. Its Web Detection module pairs typical computer vision operations with Google’s search mechanism, partly performing as a more specialized batch reverse image search engine. This feature thus allows to potentially retrieve images’ spread across the web. We discuss the implications of this non-verbal methodological approach by tracking the 'live' cross-platform circulation of images shared on Twitter in the context of 2018 FIFA World Cup Final Draw ceremony, held on December 2017. Following a novel methodological protocol, we ran several iterations of Vision API processing, thus generating a time series of URLs pointing to pages in which images matching the ones being processed were found. The study analyses in depth the circulation of four popular images about the broadcasted media event, observing their ‘live’ spreading dynamics as well as the computer vision API performance. Among the findings, we point out the adoption of images as 'traffic tags' for cross-platform analysis as a promising approach to study web circulation beyond language barriers and mainstream platforms. Also, we find relevant data to discuss the specificities of the API’s algorithms and its opacity as inherent issues of the digital methods approach.

An Innovative Smartphone-based Solution for Traffic Rule Violation Detection

This paper introduces a novel smartphone-based solution to detect different traffic rule violations using a variety of computer vision and networking technologies. We propose the use of smartphones as participatory sensors via their cameras to detect the moving and stationary objects (e.g., cars and lane markers) and understand the resulting driving and traffic violation of each object. We propose novel framework which uses a fast in-mobile traffic violation detector for rapid detection of traffic rule violation. After that, the smartphone transmits the data to the cloud where more powerful computer vision and machine learning operations are used to detect the traffic violation with a higher accuracy. We show that the proposed framework detection is very accurate by combining a) a Haar-like feature cascade detector at the in-mobile level, and b) a deep learning-based classifier, and support-vector machine-based classifiers in the cloud. The accuracy of the deep convolutional network is about 92% for true positive and 95% for true negative. The proposed framework demonstrates a potential for mobile-based traffic violation detection by especially by combining the information of accurate relative position and relative speed. Finally, we propose a real-time scheduling scheme in order to optimize the use of battery and real-time bandwidth of the users given partially known navigation information among the different users in the network, which us the real case. We show that the navigation information is very important in order to better utilize the battery and bandwidth for each user for a small number of users compared to the navigation trajectory length. That is, the utilization of the resources is directly related to the number of available participants, and the accuracy of navigation information.

Efficient solution of Otsu multilevel image thresholding: A comparative study

Multi-level thresholding of a gray image is one of the basic operations in computer vision, with applications in image enhancement and segmentation. Various criteria for the selection of threshold level values were proposed. One of these criterion is the Otsu criterion that uses maximization of between-class variance approach. Although applying multi-level thresholding to an image is a straightforward operation, computation of the threshold levels with Otsu criterion is a computationally expensive process. In this paper, we revisit a dynamic programming algorithm that provides exact and efficient solution to the problem and compare it with modern meta-heuristic algorithms. We provide a rigorous proof for the correctness of the algorithm. The algorithm computational cost is linear in the number of threshold levels. We compare the algorithm with state of the art algorithms and verify its superior performance. The experiments show that we could gain speedup up to 2.45 × .

Image segmentation based on adaptive K-means algorithm

Image segmentation is an important preprocessing operation in image recognition and computer vision. This paper proposes an adaptive K-means image segmentation method, which generates accurate segmentation results with simple operation and avoids the interactive input of K value. This method transforms the color space of images into LAB color space firstly. And the value of luminance components is set to a particular value, in order to reduce the effect of light on image segmentation. Then, the equivalent relation between K values and the number of connected domains after setting threshold is used to segment the image adaptively. After morphological processing, maximum connected domain extraction and matching with the original image, the final segmentation results are obtained. Experiments proof that the method proposed in this paper is not only simple but also accurate and effective.

Recurrently Decomposable 2-D Convolvers for FPGA-Based Digital Image Processing

Two-dimensional (2-D) convolution is a widely used operation in image processing and computer vision, characterized by intensive computation and frequent memory accesses. Previous efforts to improve the performance of field-programmable gate array (FPGA) convolvers focused on the design of buffering schemes and on minimizing the use of multipliers. A recently proposed recurrently decomposable (RD) filter design method can reduce the computational complexity of 2-D convolutions by splitting the convolution between an image and a large mask into a sequence of convolutions using several smaller masks. This brief explores how to efficiently implement RD-based 2-D convolvers using FPGA. Three FPGA architectures are proposed based on RD filters, each with a different buffering scheme. The conclusion is that RD-based architectures achieve higher area efficiency than other previously reported state-of-the-art methods, especially for larger convolution masks. An area efficiency metric is also suggested, which allows the most appropriate architecture to be selected.

APPLYING CHAOTIC IMPERIALIST COMPETITIVE ALGORITHM FOR MULTI-LEVEL IMAGE THRESHOLDING BASED ON KAPUR’S ENTROPY

Segmentation is one of the most important operations in image processing and computer vision. Normally, all image processing and computer vision applications are related to segmentation as a pre-processing phase. Image thresholding is one of the most useful methods for image segmentation. Various methods have been represented for image thresholding. One method is Kapur thresholding, which is based on maximizing entropy criterion. In this study, a new meta-heuristic algorithm based on imperialist competition algorithm was proposed for multi-level thresholding based on Kapur's entropy. Also, imperialist competitive algorithm is combined with chaotic functions to enhance search potency in problem space. The results of the proposed method have been compared with particle optimization algorithm and genetic algorithm. The find ings revealed that the proposed method was superior to other methods.

Linear Feature Detectionin Images

Linear feature detection in digital images is an important low-level operationin computer vision that has many applications. In remote sensing tasks, it can be usedto extract roads, railroads, and rivers from satellite or low-resolution aerialimages,which can be used for the capture or update of data for geographic information andnavigation systems. In addition, it is useful in medical imaging for the extraction ofblood vessels from an X-ray angiography or the bones in the skull from a CT or MRimage. It also can be applied in horticulture for underground plant root detection inminirhizotron images.

Segmentation Methods for Images: A Systematic Review

Image segmentation is the initial operation in image processing and computer vision to extract a region of interest from a still image or a sequence of images (video). Segmentation is the key operation to perform a number of applications today. Different supervised and unsupervised evaluation methods have been used for segmentation for numerous applications, but still it is a challenging task to identify which method is more accurate. This paper presents a survey on the analysis and comparison of these evaluation methods for evaluating the effectiveness of segmentation methods. It seems that more new techniques have been obtained in these years that are useful for improving accuracy, but still, it is an area of on-going investigation. Keywords: edge detection, evaluation, segmentation, thresholding, histogram