| Title |
Classification of Emotional Adjective for the Hospital Indoor Image Based on Deep Learning |
| Authors |
정민영(Chung, Minyoung) ; 이현수(Lee, Hyunsoo) |
| DOI |
https://doi.org/10.14774/JKIID.2019.28.6.075 |
| Keywords |
Classification; Emotional adjective; Hospital; Indoor image; Deep learning |
| Abstract |
This research suggests to classify Emotional adjective for the hospital indoor image. The aim of this study was twofold. First it is an attempt to overcome limitation of overfitting data with pre-process and Second it is an approach for prediction quantified the image data with Emotional adjective. The emotion is important because emotion interact between indoor and human. The hospital indoor image also have specialized emotional effect.
Emotional adjective is necessary to verify throughout variety of source qualitative and quantitative research.
recently it is getting more harder with I.R.B.(Institutional Review Board) than Emotional adjective data had made.
This research is based on deep learning method for emotional adjective quantifiaction that can replace thousands of people’s cognition. In the proposed simulation, emotional colors are firstly processed in the frequency domain to indoor images which can be treated as an emotional image. For pre-processing Emotional colors are extracted from hospital image. and search the emotional adjetive to get indoor images to fed in CNN(Convolutional Neural Network). For the hospital indoor image clustered, emotional indoor image are fed in CNN. The output of the CNNs are fused using TF(TensorFlow) API. The input of the fusion is given to a support of Python language for image classification. The proposed system is evaluated using Tensor board - which is the proved data. This research has concluded that it is desirable to use TF for predicting the set of emotional adjective and it helps for emotion analysis efficiently. TF works for the emotional image classifying the hospital indoor images. The hospital image is classified using deep learning, and analysis of emotion as A is 80 percentage modern and B is 20 percent natural in a second for a thousand emotional colors. It is expected to use these results of research have for implications of emotional analysis that represent functions of the indoor images. |