Research And Scientific Works

In this paper, we present several approaches to configuration of deep convolutional neural networks for image classification. A common problem when creating deep structures is their proper designing and configuration. This paper shows the learning of the baseline model for image classification and its variations with different structures based on the baseline model. Each of them has different configurations related to downsampling, pooling and filters dilatation. The paper is intended as a guideline for proper designing of deep structures based on experiences resulting from the modifications of deep models configurations.

International Conference on Artificial Intelligence and Soft Computing / 2019 / Springer, Cham / p. ( 223-235)

In the paper, a neuro-fuzzy structure is implemented as a movie recommender. First, a novel method for transforming nominal values of attributes into a numerical form is proposed. This allows representing the nominal values, e.g. movie genres or actors, in a neuro-fuzzy system designed from scratch using the Mendel-Wang algorithm for rules generation. Several experiments illustrate performance of the neuro-fuzzy recommender.

International Conference on Artificial Intelligence and Soft Computing / 2018 / Springer, Cham / p. ( 752-762)

In this paper, we present our novel approach to recommender systems based on a neuro-fuzzy approach. The neuro-fuzzy approach allows for deciding to recommend or not to recommend processed items for a user. By using it, we can understand the decision through analyzing rules of decision paths. Our method gives a possibility to learn and simulate users decisions based on their actions in our test environment. Finally, a rank list of top-rated items is delivered to the user based on simulated rank for each of them. We develop our AI framework to perform tests with the use of CUDA technology. Additionally, we develop a user interface in the form of a web application. It gives the possibility to perform simulations of real users. To compare our approach with a deep learning based method, we perform tests on the MovieLens 20M Dataset. It should be noted that the architecture of the data module of our system ...

2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) / 2018 / IEEE

In the case of building large convolutional neural networks, signal propagation speed is one of priority factors. Training large neural structures requires enormous time for achieving satisfying accuracy. In addition, the networks need to be learn by very large sets of good quality training images, which is another time-consuming factor. The paper presents a fast computing framework with some methods to optimize the signal propagation speed. We compare our implementation with the original OverFeat implementation.

2016 IEEE International Conferences on Big Data and Cloud Computing (BDCloud), Social Computing and Networking (SocialCom), Sustainable Computing and Communications (SustainCom)(BDCloud-SocialCom-SustainCom) / 2016 / IEEE / p. ( 118-123)

In this paper we present a system intended for content-based image retrieval tightly integrated with a relational database management system. Users can send query images over the appropriate web service channel or construct database queries locally. The presented framework analyses the query image based on descriptors which are generated by the bag-of-features algorithm and local interest points. The system returns the sequence of similar images with a similarity level to the query image. The software was implemented in .NET technology and Microsoft SQL Server 2012. The modular construction allows to customize the system functionality to client needs but it is especially dedicated to business applications. Important advantage of the presented approach is the support by SOA (Service-Oriented Architecture), which allows to use the system in a remote way. It is possible to build software which ...

International Conference on Artificial Intelligence and Soft Computing / 2016 / Springer, Cham / p. ( 746-754)

Detection of bone area in digital X-ray images and methods of comparing such images on the basis on their content is still an issue which can be substantially improved. In this paper we present a new method of efficient bone identification and its description by a collection of simple geometrical shapes. The idea of this kind of bones description was to developed a method that could reduce the amount of data to minimum. The solution enables fast comparison of X-Ray images by checking small amount of data. This kind of geometric description of bone area is designed to create a robust bone descriptor which will be used as image pattern for image comparison method. The assumption is to create a descriptor of X-ray digital image content, mining them in large databases and search and compare X-ray images on the basis of their content. The achievement of the objectives was possible through the use of an edge ...

2015 IEEE Symposium Series on Computational Intelligence / 2015 / IEEE / p. ( 1337-1342)

This paper describes a method for searching for common sets of descriptors between collections of images. The presented method operates on local interest keypoints, which are generated using the SURF algorithm. The use of a dictionary of descriptors allowed achieving good performance of the content-based image retrieval. The method can be used to initially determine a set of similar pairs of keypoints between images. For this purpose, we use a certain level of tolerance between values of descriptors, as values of feature descriptors are almost never equal but similar between different images. After that, the method compares the structure of rotation and location of interest points in one image with the point structure in other images. Thus, we were able to find similar areas in images and determine the level of similarity between them, even when images contain different scenes.

International Conference on Artificial Intelligence and Soft Computing / 2015 / Springer, Cham / p. ( 747-756)

A framework for detecting loss of consciousness and epilepsy attack based on a neuro-fuzzy system embedded in an accelerometer built-in mobile phone is presented. Additional filtering algorithms protect the system against excessive energy consumption. The system has the ability to monitor and control daily user behaviour as well as to react to situations that can be life or health threatening, with a self-learning mechanism that can adjust to motility of human movement. Moreover, an advantage of our system, is a function of quick contact with appropriate services or relatives, by sending health state and location data regarding the person, in case the user loses consciousness or has an epilepsy seizure.

International Conference on Artificial Intelligence and Soft Computing / 2015 / Springer, Cham / p. ( 142-150)

This paper presents a novel relational database architecture aimed to visual objects classification and retrieval. The framework is based on the bag-of-features image representation model combined with the Support Vector Machine classification and is integrated in a Microsoft SQL Server database.

2015 IEEE 2nd International Conference on Cybernetics (CYBCONF) / 2015 / IEEE / p. ( 478-482)

This paper describes a novel method of image key-point descriptor indexing and comparison used to speed up the process of content-based image retrieval as the main advantage of the dictionary-based representation is faster comparison of image descriptors sets in contrast to the standard list representation. The proposed method of descriptor representation allows to avoid initial learning process, and can be adjusted taking into consideration new examples. The presented method sorts and groups components of descriptors in the process of the dictionary creation. The ordered structure of the descriptors dictionary is well suited for quick comparison of images by comparing their dictionaries of descriptors or by comparing individual descriptors with the dictionary. This allows to skip a large part of operations during descriptors comparison between two images. In contrast to the standard dictionary, our method takes ...

2014 International Joint Conference on Neural Networks (IJCNN) / 2014 / IEEE / p. ( 512-517)

In this paper we present a new algorithm for translating visual information into a semantic form. In our approach we try to combine these two separate areas of computer since into one process. The main goal is to achieve very good performance at searching for similar images. In this paper we explain in details the design of the translation algorithm which is only one part of the whole process, but the most important one. This module is some kind of interface between information in the form of digital image and the information represented by lexems. We will also concisely demonstrate the structure of the whole SIA (Semantic Image Analysis) project.

International Conference on Artificial Intelligence and Soft Computing / 2014 / Springer, Cham / p. ( 783-792)

This paper presents a concept of an object pre-classification method based on image keypoints generated by the SURF algorithm. For this purpose, the method uses keypoints histograms for image serialization and next histograms tree representation to speed-up the comparison process. Presented method generates histograms for each image based on localization of generated keypoints. Each histogram contains 72 values computed from keypoints that correspond to sectors that slice the entire image. Sectors divide image in radial direction form center points of objects that are the subject of classification. Generated histograms allow to store information of the object shape and also allow to compare shapes efficiently by determining the deviation between histograms. Moreover, a tree structure generated from a set of image histograms allows to further speed up process of image comparison. In this ...

International Conference on Artificial Intelligence and Soft Computing / 2014 / Springer, Cham / p. ( 639-650)

Segmentation of digital images is an important issue of object recognition. This method of image processing allows to determine single object areas in images. This paper presents an improved segmentation method which gives a possibility to detect single objects in images by using the disparity map algorithm in connection with the mean shift pixel grouping algorithm. Images are processed in grayscale where range of colors is in from 0 to 255. Grayscale allows to detect objects on the basis of pixels brightness. To achieve this purpose we used one of grouping algorithms known as mean shift. Images obtained from mean shift are in the form of separated images which could be subject of further processing. Important feature of mean shift processing is that we obtain the results in the form of backgroundless images containing important objects from the input image.

International Conference on Parallel Processing and Applied Mathematics / 2013 / Springer, Berlin, Heidelberg / p. ( 433-443)

Digital X-ray imaging is a source of generous information about health of patient bones. One of major obstacles in computer analysis of digital X-ray images is the presence of bone tissue and soft-tissue areas. It has a negative impact on the quality of bone edge detection or detection of bones area on X-ray images. The main goal is to create an efficient method of edge detection which performs efficiently on properly prepared digital X-ray images. This paper describes a new method of background removal from X-ray images where the background is in the form of soft-tissue. The aim of this is to prepare the image to the next step of processing. We also present a new approach to edge detection of bones on X-ray images. Performance of the proposed method is achieved by eliminating unnecessary areas of the image which are not bone tissue and which are not the main region of interest. Additionally, the ...

International Conference on Artificial Intelligence and Soft Computing / 2013 / Springer, Berlin, Heidelberg / p. ( 309-319)

When we think about images, we usually think about that what we can detect by our eyes. It is easy for us, because all of the hard work is already done by our own brain. Human brain extracts from images all information which is currently important. It is not possible to mirror the whole natural process, because now we do not posses enough knowledge about our brain. Nevertheless, a lot of research is devoted to achieve even part of the targets. This is a small steps strategy, so we are not able to do all at once, but we try to test different approaches, combine and develop new digital images processing algorithms. In this paper we present a DOE (Density of Edges) algorithm and its application as a basis for the GrubCut algorithm. We also present the whole preprocessing approach and which algorithms were used. Results of that work will be used and integrated in SIA Semantic Image Analysis project ...

International Conference on Artificial Intelligence and Soft Computing / 2013 / Springer, Berlin, Heidelberg / p. ( 613-623)

This paper describes a concept of image retrieval method based on graph theory, used to speed up the process of edge detection and to represent results in more efficient way. We assume that result representation of edge detection based on graph theory is more efficient than standard map-based representation. Advantages of graph-based representation are direct access to edge nodes of the shape without search and segmentation of edges points as is the case with map-based representations. Another advance is less data consumption, only data for nodes and their connections are needed, what is important in large database applications for good scalability. In the described approach we reduce the amount of necessary image data to examine by modifying some standard edge detection method. To obtain that, we use an auxiliary grid to detect points of edge intersections with grid lines. Each ...

International Conference on Artificial Intelligence and Soft Computing / 2013 / Springer, Berlin, Heidelberg / p. ( 588-601)

This paper describes a novel image retrieval method for parasite detection based on the analysis of digital images captured by the camera from a microscope. In our approach we use several image processing methods to find known parasite shapes. At first, we use an edge detection method with edge representation by vectors. The next step consists in clustering edges fragments by their normal vectors and positions. Then grouped edges fragments are used to perform elliptical or circular shapes fitting as they resemble most parasite forms. This approach is invariant from rotation of parasites eggs or the analyzed sample. It is also invariant to scale of digital images and it is robust to overlapping shapes of parasites eggs thanks to the ability to reconstructing elliptical or other symmetric shapes that represent the eggs of parasites. With this solution we can also reconstruct incomplete shape of parasite egg ...

International Conference on Artificial Intelligence and Soft Computing / 2012 / Springer, Berlin, Heidelberg / p. ( 551-558)

In the world of computer imaging, we still do not have a good and fast enough method for image searching. This is because science is still not able to imitate fully functions of the human brain. When humans think about images, they do not think about mathematical formulas, matrices, histograms etc. Those mathematical and algorithmic methods are very good for e.g. computer face detection or number plate recognition, but we cannot directly use them for analyzing a whole image and for searching in a set of thousands or even millions of images. On the other hand, computers are able to scan millions of documents, searching for some phrase or even a single word. Fast text search is fully supported by a majority of significant database systems such as Oracle, PostgreSQL or MS SQL Server. The paper presents fast text search engine from another point of view, that is, its application in content based image ...

International Conference on Artificial Intelligence and Soft Computing / 2012 / Springer, Berlin, Heidelberg / p. ( 592-599)