Autors: Ivanova, M. S.
Title: Classification of Hand-drawn Images in Frequency Domain through Deep Learning in Support of Electronic Circuits Design
Keywords: Automation, Hand-drawn graphics, Deep learning, ResNet50

Abstract: Electronic circuit design concerns the process of finding the most suitable circuit topology according to predefined specification with circuit parameters and characteristics. Currently, the designer is supported through CAD software that proposes automation of a wide variety of engineering tasks. Anyway, such software could be improved with adoption of machine and deep learning algorithms. In this work, a method for assisting the designer is proposed. It consists of applying the ResNet50 algorithm for recognition and classification of hand-drawn graphics with amplitude-frequency response. Such an approach allows to reduce the time and effort of the designer, narrowing down the possible design variants, guiding the designer and getting him/her closer to the most suitable option. The method is verified through supporting the design of active filters. Created deep learning models are evaluated and it is proved their suitability at conductance of classification tasks.

References

    Issue

    4th Winter IFSA Conference on Automation, Robotics & Communications for Industry 4.0/5.0 (ARCI' 2024), pp. 217-220, 2024, Austria, International Frequency Sensor Association (IFSA) Publishing, S. L., DOI ISBN: 978-84-09-58219-8/ISSN: 2938-4796

    Copyright International Frequency Sensor Association (IFSA) Publishing, S. L.

    Вид: публикация в международен форум, публикация в реферирано издание, индексирана в Google Scholar