Michal Podpora

Implementation of an inexpensive EEG headset for the pattern recognition purpose

There are many types of bio-signals with various control application prospects. In this work possible control application domain of electroencephalographic signal obtained from an easily available, inexpensive EEG headset - Emotiv EPOC was presented. This work also involved application of an embedded system platform. That solution caused limits in choosing an appropriate signal processing method, as embedded platforms characterise with a little efficiency and low computing power. Potential implementation of the embedded platform enables to extend the possible future application of the proposed BCI. It also gives more flexibility, as the platform is able to simulate various environments. In this work traditional, statistical methods were neither used nor described.

Game controller based on biomedical signals

Modern, based on biomedical signals interfaces have become recently very complex, however the complexity does not always lead to increased functionality or usability. In particular, when it comes to handicapped users, the currently available solutions are far from satisfactory. In this paper an innovative approach for biomedical signals based interfaced with the implementation of an inexpensive gaming headset Emotiv EPOC was presented. The main goal was to design, and develop an intuitive and user-friendly interface based on implementation of various biomedical signals such as EEG or EMG. The project was primarily intended for handicapped users as a replacement for traditional interfaces such as keyboard or mouse, however its potential use was extended. The proposed system differs from the already existing interfaces mainly because of its versatility to work with various biomedical signals, thus enabling a single interface to be controlled with different devices. Initial investigation has proven that Emotiv EPOC headset could be applied as an inexpensive, easily available on the open market tool for HumanComputer Interaction (HCI) systems for the gaming purpose.

YUV vs RGB - Choosing a Color Space for Human-Machine Interaction

This paper describes and compares two color spaces - YUV and RGB, taking into account possible human-computer interaction applications. Human perception-oriented properties are compared, including not only file size or bandwidth, but also subjective visibility of artifacts. 1700 tests on a group of 170 people were performed to describe the subjective quality of compressed YUV and RGB images. The paper shows that the use of the YUV color space for a machine vision implementation can give better subjective image quality than the RGB color space. The authors conclude that YUV is better for machine vision implementations than RGB due to the perceptual similarities to the human vision.

Biologically reasoned point-of-interest image compression for mobile robots

In this article author describes image compression based on the idea of biological dasiayellow spotpsila - the quality/resolution is variable, depending on the distance from point-of-interest. Reducing the amount of data in robotpsilas vision system enables possibilities of using a computer cluster for non-time-critical dasiamentalpsila processing tasks like dasiamemoriespsila or dasiaassociationspsila. This approach might be especially useful in HTM-based processing of robotpsilas vision system data.

Extending vision understanding capabilities of NAO robot by connecting it to a remote computational resource

In this article authors briefly describe the concept and an exemplary implementation of a scalable vision understanding system. In the practical implementation a popular humanoid robot NAO is being used. The robots video processing algorithm is capable of connecting to a remote computational resource (server or cluster) for running more sophisticated image understanding algorithms. The novelty of this paper involves the possibility to connect to a remote computational resource if needed. An exemplary way of using such a system - by building a remote cognitive system - is given. Basic outlines for further development of new service (Knowledge As a Service) for mobile robots are also given.

Policy-based self-configuration of autonomous systems information inputs

In this paper the idea of Policy-based Input Selector, which is a framework of a highly flexible modular software interface was presented. The Policy-based Input Selector allows run-time context-sensitive and self-acting reconfiguration of information inputs in cognitive systems. The proposed solution has great application capability in autonomous and intelligent systems with multiple data inputs, such as - inter alia - vision, Laser Range Finders (LRF) or audio. Potential implementation of policies enables efficient context switching of information input and this was in detail presented and explained in this work.

Making Eye Contact with a Robot - Exploring User Experience in Interacting with Pepper

In this paper authors describe a new approach to analysis of user experience in interacting with a humanoid robot Pepper. The designers and engineers of humanoid robots struggle to make the robots appear friendly and welcoming – not only to surpass the so-called Uncanny Valley, but also to make the human-machine interaction consistent and comfortable. The paper includes description of an experiment involving humanoid robot Pepper and human volunteers in a form of a short verbal interaction. The result of the experiment consists not only of a questionnaire, but also other parameters measured and calculated by a robot, inter alia the time of the eye contact. The authors evaluate the correlation between the time of eye contact and the overall opinion of volunteers.

Initial Study on Cancer-Related Cognitive Dysfunction With the Implementation of QEEG

Abstract Chemotherapy is one of the most common treatments used in cancer therapy despite its serious side effects, which remain a huge concern. To the most common side effects (apart from weight and hair loss) include memory and concentration problems as well as changes in taste. This paper presents a very preliminary study (with only one elderly female subject participating) focusing on cognitive dysfunction after chemotherapy in breast cancer. This phenomenon, called Cancer-Related Cognitive Dysfunction (CRCD), is a frequent occurrence. The obtained results prove the impact of chemotherapy on the participant’s ability to concentrate. The QEEG method and digital data analysis were used as the measurement methods.

Innovative approach in signal processing of electromyography signals

with its potential implementation for the purpose of HMI systems, where embedded platforms are applied. The method does not involve any traditional, statistical signal processing methods. Materials and methods. The proposed method differs from the traditional signal processing methods due to the no need of using equipment with high-computing power, which results with its wide potential implementation. Signal processing of various bio-signals is currently a very dynamically developing scientific area. The innovation of the proposed solution relies on its simplicity, efficiency and waht’s more – it does not implement any statistical signal processing. Results. The proposed method has prospective implementation for the control purpose in order to improve quality of life for handicapped users. Conducted research was intended for potential application on an embedded system platform, which has caused some significant limits in choosing an appropriate signal processing method. Traditional, sophisticated, statistical signal processing methods were not used for the purpose of this work. Discussion. This paper is a pilot study for the prospective EMG-based control of an artificial hand. The conducted analysis was done in an off-line mode, however further plans on on-line signal processing plans were made. All calculations were done in MATLAB. Conclusions. The paper presents an innovation as no other similar methods were found in literature. The paper also shows

Dynamic re-definition of Region-of-Interest in vision system's feedback

In this paper the author describes a modification of the Region-of-Interest idea, in which the “best quality” area is defined individually for every frame of video stream. The robots vision system decides which fragment of the whole image should be transmitted using the best quality.