Sleep apnea syndrome (SAS) is prevalent in individuals and recently, there are many studies focus on using simple and efficient methods for SAS detection instead of polysomnography. However, not much work has been done on using nonlinear behavior of the electroencephalogram (EEG) signals. The purpose of this study is to find a novel and simpler method for detecting apnea patients and to quantify nonlinear characteristics of the sleep apnea. 30 min EEG scaling exponents that quantify power-law correlations were computed using detrended fluctuation analysis (DFA) and compared between six SAS and six healthy subjects during sleep. The mean scaling exponents were calculated every 30 s and 360 control values and 360 apnea values were obtained. These values were compared between the two groups and support vector machine (SVM) was used to classify apnea patients. Significant difference was found between EEG scaling exponents of the two groups (p < 0.001). SVM was used and obtained high and consistent recognition rate: average classification accuracy reached 95.1% corresponding to the sensitivity 93.2% and specificity 98.6%. DFA of EEG is an efficient and practicable method and is helpful clinically in diagnosis of sleep apnea.

Music genre recognition (MGR) plays a fundamental role in the context of music indexing and retrieval. Unlike images, music genres consist of immediate characteristics that are highly diversified with abstractions in different levels. However, most representation learning methods for MGR focus on global features and make decisions from features in the same level. To remedy such defects, we intergrate a convolutional neural network (CNN) with NetVLAD and self-attention to capture the local information across levels and learn their long-term dependencies. A meta classifier is used to make the final MGR classification by learning from aggregated high-level features from different local feature coding networks. Experimental results show that the proposed approach yields higher accuracies than other state-of-the-art models on GTZAN, ISMIR2004, and Extended Ballroom dataset.