Neural Network Development Based on Knowledge about Environmental Influence
| Autor: | Kirill Akhmetzyanov, Andrey N. Kokoulin, Alexander A. Yuzhakov |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Class (computer programming)
Training set Artificial neural network Computer science business.industry 05 social sciences 010501 environmental sciences Object (computer science) Machine learning computer.software_genre 01 natural sciences Convolutional neural network Image (mathematics) Development (topology) 0502 economics and business Artificial intelligence 050207 economics business computer 0105 earth and related environmental sciences |
| Zdroj: | 2020 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering (EIConRus). |
| DOI: | 10.1109/eiconrus49466.2020.9039226 |
| Popis: | The article proposes an approach to the development of the original neural network, as well as the implementation of such a network. The approach is to extract the effects that the environment has on any object. Further, the proposed neural network receives an image of the object without any impact. To train the proposed neural network, which eliminates environmental influences and compares objects without impacts, only a few training images for each recognizable class are needed, which will eliminate the lack of convolutional neural networks - the need for a large training set for accurate recognition. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|