SAMI 2017: IEEE 15th International Symposium on Applied Machine Intelligence and Informatics (Konferencia helye, ideje: Herlany, Szlovákia, 2017.01.26-2017.01.28.)

Development of Keyword-based Image Retrieval System in Automatically Annotated Image Database

András Placskó, Szabolcs Sergyán

The implemented system automatically capable of annotating images of a database based on samples. The system creates annotations to images with detection of the object belonging to the keyword. The set of keywords is pre-defined. After the user searching the first twelve relevant results will be showed. Due to the higher accuracy of the system, the user has to comment results then the system has to learn from the comment of the user.

Detection of Oesophagus Tissue Changes Using Image Processing

István Golarits, Máté István Tóth, Zoltán Vámossy

The computer-based cell and tissue analysis has always been an active area of healthcare informatics, but the popularity has greatly increased since the evolution of hardware. The process began about a decade ago, and since then it has become a key research area in medical informatics. Over the years various approaches were attempted by the developers, these are mostly based on the properties of the images, for example colors, shapes, contours and textures, and conclusions are drawn from these parameters. This topic gained popularity because medical imaging devices became more easily accessible, and they are able to produce far better quality images. The article describes a prototype system that may accelerate the detection of the signs of oesophagus adenocarcinoma, thus helping the task of pathologists and specialist in the field of diagnosis. The presence of Barrett Metaplasia greatly increases the risk of developing adenocarcinoma, therefore the developed software looks for the signs of this disease. In order to accomplish this goal, the developed image processing system evaluates tissue images produced by virtual microscopes.

Novel Method for Quadcopter Controlling Using Nonlinear Adaptive Control Based on Robust Fixed Point Transformation Phenomena

Bence G. Czakó, Krisztián Kósi

In the past few years quadcopters have become an integral part of life. There are many applications for these objects including entertaining purpose, agricultural monitoring, or gathering information from places what humans can not reach. An interesting problem is controlling these vehicles from remote locations in order to follow a desired trajectory. While many solutions exist to this task the vast majority of them heavily rely on linear control methods which are susceptible to parameter uncertainties or outer disturbances, which affect the tracking capability of the quadcopter adversely. In this paper a novel nonlinear approach, called Robust Fixed Point Transformation based adaptive control (RFPT) is presented which can provide remedy to the above mentioned disturbances.

E-assessment using image processing in ∞Exams

Ádám Tóth, István Vajda, Zoltán Vámossy

This paper features a software system called ∞Exams (InfinityExams) which supports (primarily in higher education) 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 approach significantly differs from the ones used in the previous 10+ years which were implemented in such a way that they could not reproduce and replace the traditional paper-based examination model.
The heart of the article relies on the most important element of the software which is the image processing flow.

A model of heat exchange and accumulation in small-sized bioreactors during ethanol fermentation

Dániel Kiss

This paper presents a mathematical model of heat production, exchange and accumulation during ethanol generator at small-scaled fermentor tanks in home wineries or breweries. Heat generation is directly related to CO2 production rate, which serves as a time-dependent input parameter of the model, as well as the actual ambient temperature. Validation at pilot scale experiments showed good agreement with recorded temperature sensor data.

Application and Properties of the Radon Transform for Object Image Matching

Gábor Kertész, Sándor Szénási, Zoltán Vámossy

The Radon transform and the reverse formulas are mostly used in computer tomography to reconstruct the original shape. The projections of an image could be used in computer vision for object recognition, based on projection matching. By comparing the projections, an effective method could be defined. To understand the projection behavior the basic properties of the transform are inspected, and the results of several image transformations are analyzed.