1. Home

VISUALIZATION METHODS OF IMAGE CLASSIFICATION PROCESS IN NEURAL NETWORKS

Kaspars Vogulis, Valdis Platonovs, Edgars Judovičs, Sergejs Kodors

Abstract


The aim of the work was to find and describe ways of visualising layers of neural netwoks, which analyse images and classify them. By visualising network layers, scientists and developers could see features, which influence on results of neural network mapping, training and overall result. In this paper, authors demonstrate different visualization methods, which can be applied by machine learning engineers.

Keywords


image classification; neural networks; layers

Full Text:

PDF

References


Chae Y.T., Horesh R., Hwang Y., Lee Y.M., Artificial neural network model for forecasting sub-hourly electricity usage in commercial buildings. [Piekļuve:12.04.2020]

Huang M., Hung Y., Lee W., Li R. K., Wang T.. Usage of Case-Based Reasoning, Neural Network and Adaptive Neuro-Fuzzy Inference System Classification Techniques in Breast Cancer Dataset Classification Diagnosis [Piekļuve:14.04.2020]

Kagaya H., Aizawa K. , Ogawa M. . Food Detection and Recognition Using Convolutional Neural Network. [Piekļuve:12.04.2020]

Zhang, H., Zou, Z., Li, J. et al. Flame image recognition of alumina rotary kiln by artificial neural network and support vector machine methods. J. Cent. South Univ. Technol. 15, 39–43 (2008). [Piekļuve:14.04.2020]

Sinha P. A symmetry perceiving adaptive neural network and facial image recognition [Piekļuve:14.04.2020]

VGG16 – Convolutional Network for Classification and Detection [Tiešsaiste] Pieejams: https://neurohive.io/en/popular-networks/vgg16/




DOI: https://doi.org/10.17770/het2020.24.6760

Refbacks

  • There are currently no refbacks.
Rēzeknes Tehnoloģiju akadēmija, 2016 © Atbrīvošanas aleja 115, Rēzekne, Latvija. Tālrunis: 28325368;
Jautājumus un ierosinājumus sūtiet uz helpdesk@rta.lv