Wyniki wyszukiwania - BazTech

1

Estimation of absolute distance and height of people based on monocular view and deep neural networks for edge devices operating in the visible and thermal spectra

Gąsienica-Józkowy Jan, Cyganek Bogusław, Knapik Mateusz, Głogowski Szymon, Przebinda Łukasz

Annals of Computer Science and Information Systems

|

2023

|

Vol. 35

503--511

EN

Accurate estimation of absolute distance and height of objects in open area conditions is a significant challenge. In this paper, we address these problems and we propose a novel approach that combines classical computer vision algorithms with modern neural network-based solutions. Our method integrates object detection, monocular depth estimation, and homography- based mapping to achieve precise and efficient estimations of absolute height and distance. The solution is implemented on the edge device, which enables real-time data processing using both visual and thermography data sources. Experimental evaluation on a height estimation dataset prepared by us demonstrates an accuracy of 97.06\% and validates the effectiveness of our approach.

2

Small footprint embedded systems paradigm based on a novel and scalable Implementation of FORTH

Cyganek Bogusław

Annals of Computer Science and Information Systems

|

2022

|

Vol. 31

123--130

EN

This paper describes architecture of the novel implementation of the Forth interpreter-compiler. The architecture follows the object- and component-oriented design paradigms. The implementation is done with the modern C++ 20 language taking full advantage of such constructs as lambda functions, variadic templates, as well as the coroutines and concepts. The system is highly modular and easily scales for small footprint embedded systems. We propose to extend Forth with the coroutine words that allow for async operations and lightweight cooperative multi-threading. We show successful deployment of the proposed Forth implementation on three platforms, two PC frameworks running Linux and Windows, respectively, as well as on tiny embedded system NodeMCU v3 with the 32-bit RISC ESP8266 microprocessor and 32/80KB memory. The project has also educational value, showing intrinsic operation of Forth and modern C++. Software is available free from the Internet.

3

New thermal automotive dataset for object detection

Balon Tomasz, Knapik Mateusz, Cyganek Bogusław

Annals of Computer Science and Information Systems

|

2022

|

Vol. 31

43--48

EN

Although there are many efficient deep learningmethods, object detection and classification in visible spectrum have many limitations especially in case of poor light conditions. To fill this gap, we created a novel thermal video database containing few thousands of frames with annotated objects acquired in far infrared thermal spectrum. Thanks to this we were able to show its usability in the traffic object recognition based on the YOLOv5 network, properly trained to gain maximal performance on thermal images, which contain many small objects and are characteristic of different properties than the visible spectrum counterparts. The proposed thermal database, as well as the fully trained model are main contributions of this paper. These are made available free for other researchers. Additionally, based on the highly efficient car detector we show its application in the car speed measurement based exclusively on thermal images. The proposed system can be also used in the Advanced DriverAssistance Systems (ADAS), and help autonomous driving.

4

Modern C++ in the era of new technologies and challenges - why and how to teach modern C++?

Cyganek Bogusław

Annals of Computer Science and Information Systems

|

2022

|

Vol. 30

35--40

EN

Computers are one of the most important inventions of the century, and computer languages enable human-computer communication. Undoubtedly, C++ is one of the most important and influential in this group. Nevertheless, new technologies and related industry challenges place high demands on C++ and foster the development of new computer languages that meet new needs. For this reason, and thanks to the dynamically operating ISO standardization group, C++ is constantly updated while maintaining its backward compatibility. However, all this complicates and hinders not only the teaching of beginners but also the use by professionals. In this article, we briefly discuss the goals as well as proposed methodologies and techniques for teaching contemporary C++ in the age of new technologies and challenges.

5

An impact of tensor-based data compression methods on deep neural network accuracy

Grabek Jakub, Cyganek Bogusław

Annals of Computer Science and Information Systems

|

2021

|

Vol. 26

3--11

EN

The emergence of the deep neural architectures greatly influenced the contemporary big data revolution. How-ever, requirements on large datasets even increased a necessity for efficient data storage. The storage problem is present at all stages, from the dataset creation up to the training and prediction stages. However, compression algorithms can significantly deteriorate the quality of data and in effect the classification models. In this article, an in-depth analysis of the influence of the tensor-based lossy data compression on the performance of the various deep neural architectures is presented. We show that the Tucker and the Tensor Train decomposition methods, with properly selected parameters, allow for very high compression ratios, while conveying enough information in the decompressed data to achieve only a negligible or very small drop in the accuracy. The measurements were performed on the popular deep neural architectures: AlexNet, ResNet, VGG, and MNASNet. We show that further augmentation of the tensor decompositions with the ZFP floating-point compression algorithm allows for finding optimal parameters and even higher compressions ratios at the same recognition accuracy. Our experiments show data compressions of 94%-97% that result in less than 1% accuracy drop.