Journal of Engineering and Applied Sciences
ISSN: Online 1818-7803ISSN: Print 1816-949x
Abstract
The following study presents the development of a color classification algorithm for convolutional neural networks and fully-connected neural networks which uses a database of 200 images per color and between 12 and 18 colors to be classified for the training of the two networks. Subsequently, a comparison was made between their accuracy percentages where the best results were 95.33% for the convolutional neural network and 93.33% for the fully connected in the recognition of 12 colors and 93.67 and 35.23% for 18 colors, respectively. Finally, the best network is selected to design a video recognition application and the results are presented.
How to cite this article:
Laura Pramparo and Robinson Jimenez Moreno. Colorimeter Using Artificial Neural Networks.
DOI: https://doi.org/10.36478/jeasci.2017.5332.5337
URL: https://www.makhillpublications.co/view-article/1816-949x/jeasci.2017.5332.5337