Ograniczanie wyników
Czasopisma help
Autorzy help
Lata help
Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 44

Liczba wyników na stronie
first rewind previous Strona / 3 next fast forward last
Wyniki wyszukiwania
Wyszukiwano:
w słowach kluczowych:  object detection
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 3 next fast forward last
EN
Deposition defects like porosity, crack and lack of fusion in additive manufacturing process is a major obstacle to commercialization of the process. Thus, metallurgical microscopy analysis has been mainly conducted to optimize process conditions by detecting and investigating the defects. However, these defect detection methods indicate a deviation from the operator’s experience. In this study, artificial intelligence based YOLOv3 of object detection algorithm was applied to avoid the human dependency. The algorithm aims to automatically find and label the defects. To enable the aim, 80 training images and 20 verification images were prepared, and they were amplified into 640 training images and 160 verification images using augmentation algorithm of rotation, movement and scale down, randomly. To evaluate the performance of the algorithm, total loss was derived as the sum of localization loss, confidence loss, and classification loss. In the training process, the total loss was 8.672 for the initial 100 sample images. However, the total loss was reduced to 5.841 after training with additional 800 images. For the verification of the proposed method, new defect images were input and then the mean Average Precision (mAP) in terms of precision and recall was 0.3795. Therefore, the detection performance with high accuracy can be applied to industry for avoiding human errors.
EN
Detection of small objects in the airspace is a crucial task in the military. In the era of today’s unmanned aerial vehicles (UAVs) technology, many military units are exposed to recognition and observation through flying objects. They are often equipped with optoelectronic warhead making a way to collect essential and secret data of the military unit. Modern technical solutions make it possible to implement some methods facilitating detection of flying objects. A lot of them utilize computer vision techniques based on image processing algorithm. Therefore, in this article, we present an analysis of the most promising algorithm for detection of small flying objects.
PL
W artykule przedstawiono analizę metod wykrywania bezzałogowych statków powietrznych wykorzystujących techniki widzenia komputerowego.
PL
W artykule opisany został problem analizy sceny na obrazach oraz sekwencjach video. Zadanie analizy sceny polega na detekcji, lokalizacji i klasyfikacji obiektów znajdujących się na obrazach. Zaimplementowany system wykorzystuje głęboką sieć neuronową, której struktura oparta została na architekturze YOLO (You Only Look Once). Niskie zapotrzebowania obliczeniowe wybranej architektury pozwala na wykonywanie detekcji w czasie rzeczywistym z zadowalającą dokładnością. W pracy przeprowadzono również badania nad doborem odpowiedniego optymalizatora wykorzystywanego w procesie uczenia. Jako przykładową aplikację wybrano analizę ruchu ulicznego w której skład wchodzi wykrywanie i lokalizacja obiektów takich jak m.in. samochody, motocykle czy sygnalizacja świetlna. Systemy tego typu mogą być wykorzystywane w wszelkiego typu systemach analizy wizyjnej otoczenia np. w pojazdach autonomicznych, systemach automatycznej analizy video z kamer przemysłowych, systemach dozoru czy analizy zdjęć satelitarnych.
EN
The paper describes the problem of scene analysis in images and video sequences. The task of scene analysis is to detect, locate and classify objects in images. As an example of an application, traffic analysis was chosen, which includes the detection and location of objects such as cars, motorcycles or traffic lights. The implemented system uses a deep neural network, whose structure is based on the YOLO (You Only Look Once) architecture. Low computing requirements of the chosen architecture allows performing real-time detection with satisfactory accuracy. The work also included a study on the selection of an appropriate optimizer used in the learning process. The program correctly detects objects with a large surface area, allowing the system to be used in traffic analysis. The work also showed that using the ADAM algorithm allowed significantly shorten the training time of the neural network. Systems of this type can be used in many types of video analysis systems such as autonomous vehicles, automatic video analysis systems with CCTV cameras, surveillance systems or satellite image analysis.
EN
Automatic image analysis is nowadays a standard method in quality control of metallic materials, especially in grain size, graphite shape and non-metallic content evaluation. Automatically prepared solutions, based on machine learning, constitute an effective and sufficiently precise tool for classification. Human-developed algorithms, on the other hand, require much more experience in preparation, but allow better control of factors affecting the final result. Both attempts were described and compared.
EN
Development of assistive robots for helping the disabled is a field of research that has gathered attention recently. According to surveys, about one billion people in the world population have some kind of disability. Dual arm robots are a suitable solution for helping people with mobility impairments. The current development in the field of dual-arm robots is focused mainly in the industrial field to carry out cyclic tasks. This includes activities such as pick and place, assembling parts and doing other industrial operations. Unlike human arms, these dual arm robots lack versatility in doing a wide variety of tasks with adequate coordination between the arms. Due to these constraints, industrial dual arm robots cannot be directly implemented for assisting the disabled. This paper focuses on designing a compact dual arm robot which closely mimics human arms to do coordinated tasks with lesser Degrees of Freedom (DoF). Therefore, the developed robot extends its capabilities from industrial applications to daily life activities. Closed loop control is used in actuating the proposed 9 DoF dual arm robot with distinct DoF. Target position is acquired using image processing. Hand to hand coordination in various operations such as pick and place, transferring objects, serving food, etc. has successfully experimented.
6
Content available remote Object detection in the police surveillance scenario
EN
Police and various security services use video analysis when investigating criminal activity. One typical scenario is the selection of object in image sequence and search for similar objects in other images. Algorithms supporting this scenario must reconcile several seemingly contradicting factors: training and detection speed, detection reliability and learning from sparse data. In the system that we propose a combined SVM/Cascade detector is used for both speed and detection reliability. In addition, object tracking and background-foreground separation algorithm together with sample synthesis is used to collect rich training data. Experiments show that the system is effective, useful and suitable for selected tasks of police surveillance.
7
Content available remote License plate detection with machine learning without using number recognition
EN
In autonomous driving, detecting vehicles together with their parts, such as a license plate is important. Many methods with using deep learning detect the license plate based on number recognition. However, there is an idea that the method using deep learning is difficult to use for autonomous driving because of the complexity in realizing deterministic verification. Therefore, development of a method that does not use deep learning (DL) has become important again. Although the authors have made the world's best performance in 2018 for Caltech data with using DL, this concept has now turned to another research without using DL. The CT5L method is the latest type, that includes techniques of the continuity of vertical and horizontal black-and-white pixel values inside the plate, unique Hough transform, only vertical and horizontal lines are detected, the top five in the order of the number of votes to ensure good performance. In this paper, a method to determine the threshold value for binarizing input by machine learning is proposed, and good results are obtained. The detection rate is improved by about 20 points in percent as compared to the fixed case. It achieves the best performance among the conventional fixed threshold method, Otsu's method, and the conventional method of JavaANPR.
8
Content available remote On effectiveness of human cell nuclei detection dependin
EN
The paper presents results of research on effectiveness of automated detection of human body cells nuclei depending on the digital image color representation used. The problem importance is presented, data representation and processing problems are discussed. The standardized machine vision-based nuclei detection procedure is proposed. Nuclei detection effectiveness measurement algorithm is presented and results are discussed. The conclusion is drawn and future work areas are indicated.
PL
W artykule przedstawiono wyniki badań skuteczności zautomatyzowanego wykrywania jąder komórkowych, w zależności od zastosowanej reprezentacji koloru przetwarzanego obrazu. Przedstawiono problemy związane z przetwarzaniem cyfrowych obrazów medycznych. Zaproponowano ujednoliconą procedurę komputerowego przetwarzania obrazu. Przedstawiono algorytm pomiaru skuteczności wykrywania jąder komórkowych w zależności od zastosowanej przestrzeni barw. Omówiono wyniki, sformułowano wnioski i wskazano przyszłe obszary badań.
PL
W artykule omówiono proces ekstrakcji parametrów morfometrycznych cyfrowych obrazów histopatologicznych na przykładzie obrazów raka piersi. Wskazano empirycznie wyznaczoną reprezentację kolorów do skutecznego zautomatyzowanego wykrywania jąder komórkowych. Przedstawiono problematykę związaną z komputerowym wspomaganiem rozpoznawania obrazów biomedycznych. Przedstawiono, zapisane w pliku csv wyniki dla przetworzonych i rozpoznanych jąder komórkowych (opis liczbowy struktur i obiektów morfologicznych). Wskazano kierunki dalszych.
EN
The paper presents process of morphometric parameters extraction of the digital biomedical image of breast cancer. There was present empirical determination of most effective color channel for automated detection of cell nuclei. The problem of computer-aided biomedical image recognition are presented. The results obtained for processed and properly recognized cell nuclei was presented. All features (numerical description of morphological structures and objects) was stored in the csv file. The future work areas are indicated.
10
EN
The article presents two methods of detecting objects in images of the surface of the earth from the air. The search was performed using local characteristic features, i.e. key points. In the first method, the corner detection was supplied using the Harris & Stephens algorithm. The descriptors were built for detection key points by the FREAK algorithm. In the second method the blobs were provided by the SURF algorithm. The descriptors were built by the SURF algorithm. After the usage of the above methods, a comparison was made. The obtained results were shown on the example images.
PL
W artykule przedstawiono dwa przykłady detekcji obiektów w zdjęciach powierzchni ziemi z powietrza. Wyszukiwanie wykonano przy użyciu cech charakterystycznych. W pierwszym przykładzie dokonano detekcji narożników przy użyciu algorytmu Harris & Stephens. Następnie zbudowano deskryptory do znalezionych punktów kluczowych w oparciu o algorytm FREAK. W drugim przykładzie zastosowano metodę SURF do odnalezienia plamek i zbudowania ich deskryptorów. Po użyciu powyższych metod dokonano porównania. Uzyskane wyniki zaprezentowano na przykładowych zdjęciach.
11
Content available Fast multispectral deep fusion networks
EN
Most current state-of-the-art computer vision algorithms use images captured by cameras, which operate in the visible spectral range as input data. Thus, image recognition systems that build on top of those algorithms can not provide acceptable recognition quality in poor lighting conditions, e.g. during nighttime. Another significant limitation of such systems is high demand for computational resources, which makes them impossible to use on low-powered embedded systems without GPU support. This work attempts to create an algorithm for pattern recognition that will consolidate data from visible and infrared spectral ranges and allow near real-time performance on embedded systems with infrared and visible sensors. First, we analyze existing methods of combining data from different spectral ranges for object detection task. Based on the analysis, an architecture of a deep convolutional neural network is proposed for the fusion of multi-spectral data. This architecture is based on the single shot multi-box detection algorithm. Comparison analysis of the proposed architecture with previously proposed solutions for the multi-spectral object detection task shows comparable or better detection accuracy with previous algorithms and significant improvement of the running time on embedded systems. This study was conducted in collaboration with Philips Lighting Research Lab and solutions based on the proposed architecture will be used in image recognition systems for the next generation of intelligent lighting systems. Thus, the main scientific outcomes of this work include an algorithm for multi-spectral pattern recognition based on convolutional neural networks, as well as a modification of detection algorithms for working on embedded systems.
12
Content available remote Object detection based on deep learning for urine sediment examination
EN
Urine sediment examination (USE) is an important topic in kidney disease analysis and it is often the prerequisite for subsequent diagnostic procedures. We propose DFPN(Feature Pyramid Network with DenseNet) method to overcome the problem of class confusion in the USE images that it is hard to be solved by baseline model which is the state-of-the-art object detection model FPN with RoIAlign pooling. We explored the importance of two parts of baseline model for the USE cell detection. First, adding attention module in the network head, and the class-specific attention module has improved mAP by 0.7 points with pretrained ImageNet model and 1.4 points with pre-trained COCO model. Next, we introduced DenseNet to the baseline model(DFPN) for cell detection in USE, so that the input of the network's head own multiple levels of semantic information, compared to the baseline model only has high-level semantic information. DFPN achieves top result with a mAP of 86.9% on USE test set after balancing between the classification loss and bounding-box regression loss, which improve 5.6 points compared to baseline model, and especially erythrocyte's AP is greatly improved from 65.4% to 93.8%, indicating class confusion has been basically resolved. And we also explore the impacts of training schedule and pretrained model. Our method is promising for the development of automated USE.
EN
The article describes recent object detection methods with their main advantages and drawbacks and shows results of application of machine learning Haar Cascade algorithm for automobile detection. The article underlines problems related to the feature dataset generation and presents an overview of current dataset augmentation methods such as image mirroring, cropping, rotating, shearing and color modification. New methods fot image dataset augmentation, such as utilization of CAD models and Deep Learning solutions, are also proposed. In order to ensure low cost, real time detection machine learning based Haar Cascade detector has been proposed and tested on a custom dataset specifically created for dataset augmentation methods evalutation. Article provides all input parameters for detector training process, along with a brief description of object detection metrics. Finally the article presents results of the baseline detector and augumented calssificator created based on vertical image mirroring technique, for different dataset configurations. Algorithms performance for real time detection on high resolution images was also evaluated.
PL
W niniejszej pracy podjęto próbę automatycznego wyodrębnienia drzew z chmury punktów na podstawie utworzonego obrazu wysokiej roślinności z przefiltrowanych danych laserowych. W tym celu został napisany skrypt w programie MATLAB. Idea jego działania opiera się na tezie, że na obrazach cyfrowych kształt drzew w górnych piętrach zbliżony jest do okręgów. Do ich detekcji posłużono się transformatą Hougha - jedną ze skutecznych metod wykrywania kształtów w widzeniu komputerowym. Badania przeprowadzono na danych pochodzących z lotniczego skaningu laserowego, obejmujących teren Cmentarza Rakowickiego w Krakowie.
EN
In the present study attempts to automatically extract trees from image which was created from points cloud representing high vegetation. For this purpose the script was written in MATLAB. The idea of the operation is based on the thesis that on the digital image trees shape in the upper floors is similar to circles. To detect trees the transform Hough was used - one of the effective methods to detect shapes in computer vision. The research was conducted on data from airborne laser scanning, which included the area of the Rakowicki cemetery in Krakow. In order to check the number of trees, a manual vectorization (indication of the trees tops) on the orthophotomap was made. However this measurement is sub-optimal, but allowed to assess the correctness of the HT algorithm.
EN
In the automated environments, mobile robots play an important role to perform different tasks such as objects transportation and material handling. In this paper, a new method for a glassy elevator handling system based on H20 mobile robots is presented to connect distributed life science laboratories in multiple floors. Various labware and tube racks have to be transported to different workstations. Locating of elevator door, entry button detection, internal buttons recognition, robot arm manipulation, current floor estimation, and elevator door status checking are the main operations to realize a successful elevator handling system. The H20 mobile robot has dual arms where each arm consists of 6 revolute joints and a gripper. The gripper has two degrees of freedom. Different sensors have been employed with the robot to handle these operations such as Intel RealSense F200 vision sensor for entry and internal buttons detection with position estimation. A pressure sensor is used for current floor estimation inside the elevator. Also, an ultrasonic proximity distance sensor is utilized for checking the elevator door status. Different strategies including HSL color representation, adaptive binary threshold, optical character recognition, and FIR smoothing filter have been employed for the elevator operations. For pressing operation, a hand camera base and a new elevator finger model are designed. The elevator finger is resolved in a way to fit the arm gripper which is used also to manipulate the labware containers. The Kinematic solution is utilized for controlling the arms’ joints. A server/client socket architecture with TCP/IP command protocol is used for data exchange between Multi-Floor System and the H20 robot arms. Many experiments were conducted in life science laboratories to validate the developed systems. Experimental results prove an efficient performance with high success rate under different lightening condition.
EN
The article presents a possible way to detect key points. The tests were carried out by the case of detection of a reference object in static images. For comparative purposes, Chris Harris & Mike Stephens [12] and Speeded-Up Robust Features (SURF) detectors [2, 3] were used. The descriptors were built based on the Fast Retina Key point (FREAK) [1, 17] and SURF algorithms [2, 3]. Six different configurations of key point detection methods with the above descriptors were implemented. The obtained results have been presented on exemplary images and in the table. They show that this type of detection of an element of interest can be successful and should be developed.
17
Content available remote Keypoint-less, heuristic application of local 3D descriptors
EN
One of the most important topics in the research concerning 3D local descriptors is computational efficiency. The state-of-the-art approach addressing this matter consists in using keypoint detectors that effectively limit the number of points for which the descriptors are computed. However, the choice of keypoints is not trivial and might have negative implications, such as the omission of relevant areas. Instead, focusing on the task of single object detection, we propose a keypoint-less approach to attention focusing in which the full scene is processed in a hierarchical manner: weaker, less rejective and faster classification methods are used as heuristics for increasingly robust descriptors, which allows to use more demanding algorithms at the top level of the hierarchy. We have developed a massively-parallel, open source object recognition framework, which we use to explore the proposed method on demanding, realistic indoor scenes, applying the full power available in modern computers.
EN
A modification of the descriptor in a human detector using Histogram of Oriented Gradients (HOG) and Support Vector Machine (SVM) is presented. The proposed modification requires inserting the values of average cell brightness resulting in the increase of the descriptor length from 3780 to 3908 values, but it is easy to compute and instantly gives ≈ 25% improvement of the miss rate at 10‒4 False Positives Per Window (FPPW). The modification has been tested on two versions of HOG-based descriptors: the classic Dalal-Triggs and the modified one, where, instead of spatial Gaussian masks for blocks, an additional central cell has been used. The proposed modification is suitable for hardware implementations of HOG-based detectors, enabling an increase of the detection accuracy or resignation from the use of some hardware-unfriendly operations, such as a spatial Gaussian mask. The results of testing its influence on the brightness changes of test images are also presented. The descriptor may be used in sensor networks equipped with hardware acceleration of image processing to detect humans in the images.
EN
An X-ray scanning and image processing have a vast range of applications in the security. An image of a content of some package being passed for example to an airplane or to the court house may help to figure out if there are any dangerous objects inside that package and to avoid possible threatening situation. As the raw X-ray images are not always easy to analyze and interpret, some image processing methods like an object detection, a frequency resolution increase or a pseudocolouring are being used. In this paper, we propose a pseudocoloring improvement over material based approach. By addition of the edge detection methods we fill and sharpen colour layers over the image, making it easier to interpret. We demonstrate the effectiveness of the methods using real data, acquired from a professional dual energy X-ray scanner.
EN
This work presents the multiscaled version of modified census features in graphical objects detection with AdaBoost cascade training algorithm. Several experiments with face detector training process demonstrate better performance of such features over ordinal census and Haar-like approaches. The possibilities to join multiscaled census and Haar features in single hybrid cascade of strong classifiers are also elaborated and tested. The high resolution example images were used in detector training process.
first rewind previous Strona / 3 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.