Learning-based Sleep Quality Evaluation

Analysis of sleep stages is an important issue for understanding optimal sleep environments. However, most studies focus on classifying sleep stages, not on sleep quality. In this work, we develop a framework to evaluate sleep quality by analyzing sleep staging patterns and defining a sleep index for quantification. By exploiting HMMs trained by reference patterns, we compute similarity measures with the structure-based method that is robust to noise. To demonstrate the validity of the proposed method, we conduct experiments using two publicly available MASS and PSG-Audio datasets.