Wyniki wyszukiwania - BazTech

1

Detecting visual objects by edge crawling

Grycuk Rafał, Wojciechowski Adam, Wei Wei, Siwocha Agnieszka

Journal of Artificial Intelligence and Soft Computing Research

|

2020

|

Vol. 10, No. 3

223--237

EN

Content-based image retrieval methods develop rapidly with a growing scale of image repositories. They are usually based on comparing and indexing some image features. We developed a new algorithm for finding objects in images by traversing their edges. Moreover, we describe the objects by histograms of local features and angles. We use such a description to retrieve similar images fast. We performed extensive experiments on three established image datasets proving the effectiveness of the proposed method.

2

Integracja danych lidarowych i fotogrametrycznych w procesie automatycznego wykrywania obiektów

Marmol U.

Archiwum Fotogrametrii, Kartografii i Teledetekcji

|

2009

|

Vol. 20

275--284

PL

Artykuł przedstawia metodę integracji danych lidarowych i fotogrametrycznych w procesie automatycznego wykrywania obiektów. Zdjęcia lotnicze stanowią klasyczną metodę pozyskiwania informacji o przestrzeni. Ostatnie lata to coraz powszechniejsze stosowanie lidaru jako źródła informacji. Zdjęcia lotnicze cechują się możliwością wykrywania granic obiektów, ale w procesie ekstrakcji cech, często dochodzi do nadmiernego oszacowania lub zaniżenia liczby obiektów. Dane lidarowe dostarczają bezpośredniej informacji o wysokości obiektów, ale posiadają ograniczenia związane z dokładnym wyznaczeniem krawędzi obiektów. Można zatem powiedzieć, że techniki przetwarzania danych: fotogrametryczna i laserowa dostarczają danych komplementarnych, a ich integracja może przyczynić się do poprawy jakości uzyskiwanych wyników. W artykule przedstawiono badania nad integracją fotogrametrii i danych laserowych w procesie wykrywania obiektów 3D – budynków i drzew. W procesie automatycznej segmentacji zostały wykorzystane cechy teksturalne pochodzące ze zdjęć lotniczych. Obiekty 3D zostały wyodrębnione na podstawie danych lidarowych, jako różnica NMPT i NMT. Przeprowadzone badania ujawniły duży potencjał danych zintegrowanych w procesie automatycznego wykrywania obiektów

EN

This paper describes a method of integrating LIDAR data and aerial images in the process of automatic object extraction. Aerial photos are classical method for obtaining spatial information. However, in recent years, LIDAR data has become more and more popular as a source of information. Aerial imagery has the ability to delineate object boundaries, but during feature extraction, the number of objects may be overestimated or underestimated. LIDAR data provide direct information about the height of an object, but have limitations when identifying boundaries. Therefore, we can say that photogrammetric sensors and LIDAR provide complementary data and their integration can improve the quality of the results. This paper presents a study of the integration of photogrammetry and LIDAR in the process of extraction of 3D objects: buildings and trees. Textural filters have been used in the automatic segmentation process. 3D objects have been separated from LIDAR data, as a DSM and DTM difference. The study has revealed the high potential and flexibility of integrated data in the automatic process of object extraction.

3

Object extration as a basic process for content-based image retrieval (CBIR) system

Jaworska T.

Opto-Electronics Review

|

2007

|

Vol. 15, No. 4

184-195

EN

This article describes the way in which image is prepared for content-based image retrieval system. Automated image extraction is crucial; especially, if we take into consideration the fact that the feature selection is still a task performed by human domain experts and represents a major stumbling block in the process of creating fully autonomous CBIR systems. Our CBIR system is dedicated to support estate agents. In the database, there are images of houses and bungalows. We put all our efforts into extracting elements from an image and finding their characteristic features in the unsupervised way. Hence, the paper presents segmentation algorithm based on a pixel colour in RGB colour space. Next, it presents the method of object extraction applied to obtain separate objects prepared for the process of introducing them into database and further recognition. Moreover, we present a novel method of texture identification which is based on wavelet transformation. Due to the fact that the majority of texture is geometrical (such as bricks and tiles) we have used the Haar wavelet. After a set of low-level features for all objects is computed, the database is stored with these features.

4

Histogram Thresholding using Beam Theory and Ambiguity Measures

Sen D., Pal S.K.

Fundamenta Informaticae

|

2007

|

Vol. 75, nr 1-4

483-504

EN

This paper presents a novel histogram thresholding technique based on the beam theory of solid mechanics and the minimization of ambiguity in information. First, a beam theory based histogram modification process is carried out. This beam theory based process considers a distance measure in order to modify the shape of the histogram. The ambiguity in the overall information given by the modified histogram is then minimized to obtain the threshold value. The ambiguity minimization is carried out using the theories of fuzzy and rough sets, where a new definition of rough entropy is presented. The applications of the proposed scheme in performing object and edge extraction in images are reported and compared with those of a few existing classical and ambiguity minimization based schemes for thresholding. Experimental results are given to demonstrate the effectiveness of the proposed method in terms of both qualitative and quantitative measures.

5

Designing pruned neighborhood neural networks for object extraction from noisy background

Bhattacharyya S., Dutta P.

Foundations of Computing and Decision Sciences

|

2006

|

Vol. 31, No. 2

105-134

EN

Refinement of neural network architectures by pruning the network interconnections reduces the computational overhead associated with the tasks for which the network is employed. A fuzzy set theoretic approach for designing pruned neighborhood topology-based neural networks for efficient extraction of objects from a noisy background, is presented in this paper. Pruning of the network architecture Is achieved by means of a judicious selection of the participating nodes of the neighborhood topology-based neural network using the fuzzy cardinality measures of the object scene. An application of the proposed methodology for designing a pruned multilayer self organizing neural network for the extraction of binary and gray scale objects from noisy backgrounds with different noise levels is demonstrated. The results obtained are compared with the outputs obtained with the conventional fully connected network architecture of the same network. Comparative results show a significant reduction in the architecture of the network with increasing noise levels for both the binary and gray scale images. Moreover, the qualities of the extracted images obtained using the pruned network architecture are found to be better than those obtained using the conventional fully connected architecture.

6

Extending DIPS software package for research and routine investigations of biomedical images

Piętka D., Dulewicz A., Jaszczak P., Nechay A.

Machine Graphics and Vision

|

1998

|

Vol. 7, No. 4

711-723

EN

The paper introduces DIPS (Digital Images Procesing System), a new, original software package, developed to meet specific requirements of image analysis in scientific environment. Without gooing deeply into programming details, key features of the system are highlighted: flexibility when designing new picture processing schemes and simplicity when performing routine tasks.