ملخص البحث :
Abstract. Happiness can be defined as a state of contentment, love, amusement, satisfaction and joy. Researchers shows
that the safe and happy work, study, or living environments can directly affect our well-being. In this article, we suggest
an algorithm to assess a voice recognition-based environment for Arabic-speakers. The proposed algorithm contains four
stages. First, the captured conversations are divided into multiples of a single voice that present one word. Second stage,
the voice of each word is converted into spectrogram image. Third stage, the spectrogram image is classified using a
Convolutional Neural Network (CNN) algorithm into positive, negative or neutral. Finally, the happiness index is
calculated based on the amount of the positive words. The suggested classifier technique can robustly categorize the
voice words with more than 63% accuracy, according to experimental results using 1000 different voiced words.
-
سنة النشر : 2024
-
تصنيف البحث : scopus
- تحميل