Vision Library Research Articles

Performance Analysis of Human Activity

This project aims to develop an AI-powered gym assistant using Jupyter Notebook and MediaPipe, a popular computer vision library, to count the repetitions of three joint exercises: curls, squats, and sit-ups. The system will provide real-time feedback and monitoring, allowing users to track their progress and improve performance. The proposed method utilizes MediaPipe, which offers pre-trained machine-learning models for human pose estimation and hand tracking. These models will accurately detect and track critical body joints and hand movements during the exercises. The system will then analyze the detected poses to identify the repetitions of each activity based on predefined movement patterns and pose thresholds. Jupyter Notebook will be used as the development environment for coding and testing the system. Python programming language and MediaPipe’s Python API will be employed to implement the pose estimation and repetition counting algorithms. The system will also incorporate a user-friendly interface, allowing users to interact with the gym assistant and receive feedback on their exercise performance. The completed project will provide an AI-powered gym assistant that can accurately count the repetitions of curls, squats, and sit-ups in real-time. Additionally, this project will contribute to the advancement of the field of fitness technology by showcasing the potential of combining computer vision and artificial intelligence techniques for gym monitoring and performance tracking. The results of this project have the potential to benefit fitness enthusiasts, trainers, and researchers alike, providing contributions to the field of fitness technology.

Heart rate measurement using image recognition technology

The measurement of heart rate (HR) has numerous applications in various fields, such as the internet of things, security, sports, and telemedicine. There are many methods for measuring pulse rates, and this research is based on a novel technique of measuring the heartbeat using image recognition technology. The innovations in the field of visual objects have made the detection process easy and quick, with high efficiency. Four step-based algorithms, including a computer, an external high-definition camera, and an open-source computer vision library, have been presented for measuring heart rate. The first step was the face detection (FD) algorithm, and the second was the area attention algorithm to determine the region of interest (ROI). The ROI signal analysis algorithm was used in the third step, using a fast Fourier transform (FFT) for frequency detection. The pulse measurement phase was the final step, and it was based on the strength of the color concentration in proportion to the time extracted from video clips. With the help of our recorded database of 50 participants based on different ages and skin colors, the process was carried out. The results of this study contributed to the development of an HR detection technique based on image recognition using the Python programming language. This is a very comfortable and effective method for measuring the human heart rate. This research article discussed various factors and obstacles that affect heart rate measurement. The results found that our system is highly competent in measuring heart rate.

Detection of Certain Objects Wearing Masks in Real Time To Prevent the Spread of the Virus (Yolov3)

A significant increase in the spread of the corona virus (COVID-19) in the community is currently happening due to people not following the health protocol rules set by the ministry of health. One of the rules is to require people to wear masks while they are outside the home. Measures need to be implemented in anticipation of situations where people do not wear masks in public spaces. Therefore, the establishment of a mask detection system is chosen as a solution in order to solve the mentioned problem above. A real-time identification system for people wearing masks is proposed to be developed in this paper. The system utilizes Yolov3 with Darknet -53 as a deep learning mask detector and OpenCV as a real-time computer vision library, so that people doing activities in a public space captured by a video can be recognized and detected when they do not wear masks . In implementing deep learning, a data set of 4000 images is divided into two classes, i.e.,2000 images with masks f or data testing purposes and another 2000 images without masks for training custom objects. The Extreme Programming (XP) method as part of the Agile Process Model is adopted for system development. Computer language support and the latest system development tools have made it possible to utilize this method in an effort to develop this system rapidly. Requirement Analysis is conducted to obtain required processes before designing system. Writing code and testing the system will be the next step before the system is declared ready to be implemented in the public space. By adopting the XP development method, all of the above steps can be implemented repeatedly until the system delivers the expected results

Design and Implementation of Chess and Card Recognition System Based on Deep Learning

In order to improve the speed and accuracy of mahjong factory packaging detection, the neural network-based chess and card recognition system designed in this paper mainly includes image preprocessing and image recognition. The image preprocessing uses the OpenCV computer vision library to segment the complete chess and cards into individual chess and cards. It is mainly to grayscale the image, use Gaussian blur to denoise the image, calculate the gradient of the image, transform the gradient image into a threshold image, perform morphological operations on the image, and perform contour detection to find the surrounding matrix of the contour. The target and the background can be divided by the vertex coordinates, and a 9x4 grid is superimposed on the divided target image and can be divided into a single mahjong chess card for neural network recognition. Image recognition uses the Vision Transformer neural network. In this paper, we first use the convolution layer to obtain the feature map, and use the feature map as the input of the Vision Transformer. The data set is mainly derived from the shooting of the mobile phone and the image expansion. Here, the image is mainly expanded according to the image brightness, blur and rotation at a certain angle. Finally, the accuracy rate of the model on the test set can reach 98%. Finally, the trained model is deployed on the android mobile terminal using TensorFlow-Lite.

COMPARISON OF METHODS FOR AUTOMATIC LICENSE NUMBER RECOGNITION

The paper is devoted to the problem of automatic detection and recognition of license plates, the solution of which has many potential applications, from security to traffic management. The purpose of this work was to compare the methods of finding and recognizing car number plates, based on the application of deep learning algorithms, which takes into account different regional standards of car number plates, video quality, different speeds of vehicles, the location of the camera in relation to the vehicle license plate, defects of the car number plate (pollution , deformation), as well as changes in external lighting conditions. The advantages and disadvantages of localization and segmentation of car number plates on cars using image binarization, Viola–Jones and Harr methods are given. It was determined that adaptive approaches are better due to the possibility of compensating the impact of obstacles on different areas of the image, for example, the distribution of shadows due to the heterogeneity of illumination. It was determined that many methods in real algorithms rely directly or indirectly on the presence of number limits. Even if the limits are not used when the number is determined, they have the possibility to be used for further analysis. The methods of templates, image histograms, and contour analysis were compared to identify familiar features in the image (segmentation). It is shown that an effective approach for recognition of car license plates can be based on the application of the methods of Viola-Jones, Harr, the analysis of brightness histograms and the SVM method. Formulated conclusions on the effectiveness of the implementation of each of the procedures were confirmed as a result of conducting experiments with the developed software in the python 3 language using the cv2 computer vision library. The described approach makes it possible to obtain a fairly high recognition accuracy at different angles of rotation of the license plate relative to the camera. Keywords: automatic recognition, license plates, localization, normalization, segmentation, character recognition.

Optimize Image Processing Algorithm on ARM Cortex-A72 and A53

This work presents a technique to optimize processing image algorithms. The increasing demand for video applications like context-aware computing on mobile embedded systems requires the use of computationally intensive image processing algorithms. The system engineer has the mandate to optimize them so as to meet real-time deadlines. A methodology to take advantage of the asymmetric dual-core processor, which includes an ARM supported by shared memory, is presented with implementation details. The target platform chosen is the NanoPi M4V2. It has a dual-core and quad-core architecture with an ARM Cortex-A72 and Cortex-A53. The basic image correlation algorithm is chosen for benchmarking as it finds widespread application for various template-matching tasks such as face recognition. The basic algorithm prototypes conform to OpenCV, a popular computer vision library. OpenCV algorithms can be easily ported to the ARM core, which runs a popular operating system such as Linux. The algorithms are tested on a variety of images, and performance results are presented, measuring the speedup obtained due to dual-core and quad-core implementation. A major advantage of this approach is that it allows the ARM processor to perform important real-time tasks.

IDENTIFICATION ALGORITHM FACES AND CRIMINAL ACTIONS

Currently, there are a number of unresolved problems in the identification of images. If a person is wearing something on their face, such as a mask or glasses, or at some point part of the face is covered by clothing, hair or an object, then the video surveillance system may lose sight of the person. Identification deteriorates significantly, and recognition of a person occurs only after some time. The purpose of this work is to improve the existing methods of recognition. The paper proposes an algorithm based on the multi-cascade method and the object detection method. This algorithm is able to identify a person by the actions of a criminal nature and by the face by highlighting some parts of the face in the form of squares and rectangles using the computer vision library. As a result of testing, the algorithm showed high detection accuracy using a GPU with 16 GB of video memory.

REAL TIME FINGER WRITING RECOGNITION USING WEBCAMERA

This paper proposes a real time finger writing recognition system using a web camera to enable user to interact with computer systems by their own bare hands. The hand region is firstly extracted by background subtraction and blob labelling . Then, convex hull and convexity defect are used to count the fingers and detect the fingertip coordinates. The trajectory of the fingertip is smoothed and drawn to become a finger writing character. This paper describes a system which uses multiple visual process to detect like Hand Tracking, Volume Gesture, Finger counter which ultimately tracks Finger Recognition based on computer vision. In modern technologies video tracking and processing the feed has been very essential. This processed data can be used for many research purposes or to express a particular output on a particular system. There are various methods for processing and manipulation of data to get the required output. This application is created using OpenCV module and Mediapipe. OpenCV is a computer vision and machine learning software library that includes many common image analysis algorithms that will help us build custom intelligent computer vision applications. In this application frequent image feed results in video tracking of our particular finger. Video tracking is the process of locating a moving object (or multiple objects) over time using a camera. The ability to perceive the shape and motion of hands can be a vital component in improving the user experience across a variety of technological domains and platforms .One such technology is using MediaPipe .MediaPipe is a framework mainly used for building audio, video, or any time series data. With the help of the MediaPipe framework, we can build very impressive pipelines for different media processing functions.

E-OMR ASSESSMENT USING IMAGEPROCESSING

E-OMR system which supports paper-based examination and makes it easier, more comfortable and speeds up the whole process while keeping every single positive attribute of it but also reducing the number of negative aspects. The way of conducting testing the knowledge of a person using Multiple Choice Questions (MCQ) has been increased gradually. In Educational industries like schools and colleges it is more common now-a-days having tests using multiple choice questions. Even in conducting interviews it is used. Current day scenario is either using OMR technology to correct the test or manually. In real-time it is quite difficult to have OMR at all the time and manually it is highly taking the time to correct and it may give you the error. We address this issue, in our proposed system we are using digital image processing technique to correct the answer using python. We are using Open Source Computer Vision Library (Open CV) to process and correct the answer. Python is the best language to implement this concept with the available Open CV library. In this system we also implement in the django environment.

FUZZY ESTIMATION PROCEDURE OF THE CONCENTRATION OF THE COMPONENTS OF A SULFURIC ACID PICKLING SOLUTION

The article is devoted to the issue of determining the level of concentration of individual components of working solutions. The relevance of the topic lies in the fact that some impurities in solutions (for example, when pickling steel) cannot be measured automatically, which reduces the efficiency of the process and increases the load on the solution regeneration section. Also, impurities are hazardous waste for the environment and humans.
 The development of a method for the operational control of the composition of the pickling solution components in pickling baths is the main task of the work. A detailed substantiation of the choice of the mathematical apparatus of the method is given. The process of fuzzy identification of the composition of the components is carried out by the number of pixels with certain digital color RGB codes from digital images of the control doses of the solution. A set of technical means for providing the method for practical implementation in the technological process is described. A prototype of a graphical user interface (GUI) for the proposed method for identifying the composition of technological liquid solutions, developed in the Processing 3.0 integrated environment using applications from the Open Source Computer Vision Library, is presented.
 A technique for extracting experimental data for filling the base of a fuzzy classifier with model reference points is proposed. Its essence lies in simultaneous measurements of the concentration of the components of the pickling solution by classical methods of titration and dry residues and the proposed method with a step equal to the regulated accuracy. The results of approbation of the proposed method are presented.
 The presented technique for obtaining a fuzzy classifier can be used to develop IT tools for the operational control of the composition of various liquid solutions with opaque colored components and color identification of defects in various flat surfaces, provided that they are color distinguishable and determine the boundaries of RGB - intervals for assessing the color of individual component (defects).

An Improved Python-Based Image Processing Algorithm for Flotation Foam Analysis

For industrial flotation foam image processing, accurate bubble size measurement and feature extraction are very important to optimize the flotation process and to improve the recovery of mineral resources. This paper presents an improved algorithm to investigate mineral flotation foam image segmentation for mineral processing. Several libraries implemented for the Python programming language are used for image enhancement and compensation, quantitative analysis of factors influencing the image segmentation accuracy, and suggestions for improvement of the flotation foam image processing. The bubble characteristics-size and morphology-and the influence of the flotation conditions on the flotation foam image are analyzed. A Python implementation of the Retinex image compensation method-region-adaptive and multiscale-is proposed to address known issues of uneven illumination and shadows affecting flotation foam images, thereby improving brightness uniformity. Finally, an improved version of the watershed segmentation algorithm included in the Python Open Source Computer Vision library is used for segmentation analysis. The accuracy of the flotation foam image segmentation is 3.3% higher than for the standard watershed algorithm and the segmentation time is 9.9% shorter.

Application of Moving Target Information Perception Technology in Intelligent Supervision System.

In order to solve the problems of inaccurate information collection, incomplete information collection, and inconsistency of collected images in traditional sports injury collection methods, an application method of moving target information perception technology in intelligent supervision system is proposed. By judging and analyzing the potential motion damage posture of the motion posture intelligent tracking images, the collected motion intelligence tracking images are judged. The intelligent tracking image matrix can make up for the shortcomings of traditional images that are not connected, complete the identification, detect potential damage in time, and take targeted preventive measures and means. Finally, according to the target detection algorithm and target tracking algorithm, combined with OpenCV computer vision library and QT image library, an intelligent video surveillance target tracking simulation system is developed. The algorithm studied in this paper is to realize the target tracking function of the intelligent video surveillance system. Through the comparison of experimental results, the design method can accurately collect damage attitude information, without calculating continuous values, and the use of three-dimensional images in the positioning process can analyze the damage attitude from multiple angles.

Automated image analysis as a tool to measure individualised growth and population structure in Chinook salmon (Oncorhynchus tshawytscha)

AbstractSelective breeding programmes depend on high‐quality measurements of phenotype and genotype with repeated individualised phenotype measurements throughout the life cycle being optimal. Recent advances in electronics and computer vision technologies offer opportunities to improve both the quality, quantity and individualisation of repeated phenotypic measurements, but remain underutilised in aquaculture breeding programmes. In this study, we compare manual measurements of phenotypic traits of Chinook salmon (Oncorhynchus tshawytscha) with digital images and an automated software analysis pipeline written in the Python® programming language using the OpenCV machine vision library. Manual measurements of length, girth and weight of passive integrated transponder‐tagged individuals were compared with image‐based measures of 738 individuals over a time span from June–December 2019. Linear regressions showed strong correlations between manual and automated measurements for fork length, girth and weight (R2 = 0.989, R2 = 0.918, R2 = 0.987, respectively). Image‐based software measurements proved powerful for tracking general population changes in growth over the study period while retaining insights about subpopulations deviating from the average (e.g. losing weight). Taken together, our study demonstrates that image‐analysis can be used to estimate fish growth traits with a high degree of precision, requires reduced labour and demonstrates that additional knowledge can be gained through tracking individuals throughout production to harvest.

Performance Optimization of Object Tracking Algorithms in OpenCV on GPUs

Machine-learning-based computer vision is increasingly versatile and being leveraged by a wide range of smart devices. Due to the limited performance/energy budget of computing units in smart devices, the careful implementation of computer vision algorithms is critical. In this paper, we analyze the performance bottleneck of two well-known computer vision algorithms for object tracking: object detection and optical flow in the Open-source Computer Vision library (OpenCV). Based on our in-depth analysis of their implementation, we found the current implementation fails to utilize Open Computing Language (OpenCL) accelerators (e.g., GPUs). Based on the analysis, we propose several optimization strategies and apply them to the OpenCL implementation of object tracking algorithms. Our evaluation results demonstrate the performance of the object detection is improved by up to 86% and the performance of the optical flow by up to 10%. We believe our optimization strategies can be applied to other computer vision algorithms implemented in OpenCL.

Face and Gender Detection Based on BP Neural Network Algorithm

Face recognition technology has a wide range of applications in real life, and many applications can be developed on the basis of face recognition, such as gender recognition, age recognition, face comparison, and beauty image decoration. In this paper, based on the computer vision library OpenCV in the python language, the face is recognized, and the mature BP neural network algorithm is used to learn the photos in the database, and finally the gender detection of the face is successfully realized.

Hidden in Plain Sight: Exploring Privacy Risks of Mobile Augmented Reality Applications

Mobile augmented reality systems are becoming increasingly common and powerful, with applications in such domains as healthcare, manufacturing, education, and more. This rise in popularity is thanks in part to the functionalities offered by commercially available vision libraries such as ARCore, Vuforia, and Google’s ML Kit; however, these libraries also give rise to the possibility of a hidden operations threat , that is, the ability of a malicious or incompetent application developer to conduct additional vision operations behind the scenes of an otherwise honest AR application without alerting the end-user. In this article, we present the privacy risks associated with the hidden operations threat and propose a framework for application development and runtime permissions targeted specifically at preventing the execution of hidden operations. We follow this with a set of experimental results, exploring the feasibility and utility of our system in differentiating between user-expectation-compliant and non-compliant AR applications during runtime testing, for which preliminary results demonstrate accuracy of up to 71%. We conclude with a discussion of open problems in the areas of software testing and privacy standards in mobile AR systems.

"Comparison of Color Classification Using Computer Vision and Deep Neural Network"

Research on artificial intelligence and machine learning is currently ongoing and is focused on real-world problems. Machine learning is used by computers to make predictions based on the provided data set or existing knowledge. The main goal of our project is to use machine learning to categorize different colors while separating CNN from computer vision. In this work, we used supervised learning to categorize different hues using a binary classification approach. Color detection is the technique of identifying a color. In this scenario, humans can recognize the hue and choose with ease. A computer, however, cannot quickly recognize color. It is challenging to get a computer to quickly detect the color. Given that, we decide to pursue this initiative. Pandas, OpenCV, and the Naive Bayes algorithm are all used in Python. Naive Bayes classifiers are models that assign category labels to issue occurrences that are represented as vectors of feature values, where the category labels are selected from a finite set. There isn't a single method for training these classifiers; rather, there is a family of algorithms built on the premise that, given a category variable, the value of one feature is independent of the value of the other feature. Open-Source Computer Vision Library OpenCV was designed to be computationally effective and with a major emphasis on real-time applications. specialized video encoding for the cloud. Panda may be a platform that runs in the cloud and provides infrastructure for encoding audio and video.

Robot intelligent grasping experimental platform combining Jetson NANO and machine vision

According to the requirements of the intelligent manufacturing experimental teaching project, this experimental platform develops a robot intelligent grasping experimental platform based on Jetson NANO and vision. The platform adopts the Jetson NANO controller based on Google open source machine learning framework TensorFlow + Keras, cooperates with the computer vision library OpenCV to develop the machine vision algorithm, and realizes the grasping and handling of materials based on a 6-DOF articulated robot, it can support open project-based experimental teaching. The experimental platform integrates the knowledge of robot control, machine vision, mechatronics and other aspects of courses, which can train students’ ability to solve complex engineering problems and stimulate students’ creative thinking. And the experimental platform has good openness. In the experimental teaching, it only needs to specify the objectives to be achieved without restricting the specific implementation methods. Students can independently design and test the algorithm program of each link in the visual positioning system through software, and deeply experience the relevant theoretical knowledge and practical methods through hands-on experiments.

AI Body Detection and Teaching System based on Mediapipe Machine Learning Platform and OpenCV Computer Vision Library

To solve the problems of low intera ctivity, high cost, large amount of data, “difficult to quantify, difficult to record, difficult to supervise, difficult to analyze” of human motion detection correction devices on the market today, we designed an intelligent device based on Mediapipe machine learning platform and OpenCV computer based on Raspberry Pi, camera and display. We designed an intelligent device for AI body detection and teaching based on Mediapipe machine learning platform and OpenCV computer vision library. By combining chip, sensor, computing platform and technology level of computer vision, speech recognition and machine learning, the device can capture human movement in real time by using camera equipment, judge the accuracy and completeness of user’s movement according to the comparison of standard movement, and give feedback to the user in real time by voice broadcast and image prompt. The test results show that the device has the advantages of low cost, simple structure, intelligence, unmanned, data and accuracy, which provides a feasible solution to further enhance the convenience and accuracy of unmanned movement teaching and rehabilitation training.

Emotion Detection Using Machine Learning in Python

In this project emotion detection using its facial expressions will be detected. It can be derived from the live feed via system’s camera or any pre-existing image available in the memory. Emotions possessed by humans can be detect by machine and has a vast scope of study in the computer vision industry upon which several research have already been done. The work has been implemented using Anaconda (Jupyter Notebook) (3.10), Open-Source Computer Vision Library (OpenCV) and NumPy. The code check the video (testing dataset) is being compared to training dataset and thus emotion is predicted. The objective of this paper is to develop a system which can analyze the image and run time video and predict the expression of the person. The study proves that this project and code is workable and produces valid results.in this project we have make change to the accuracy of the running project by using the different models of python and deep learning.