No abstract

No abstract

No abstract

Automatic detection systems are important in passive acoustic monitoring (PAM) systems, as these record large amounts of audio data which are infeasible for humans to evaluate manually. In this paper we evaluated methods for compensating class imbalance for deep-learning based automatic detection of acoustic chimpanzee calls. The prevalence of chimpanzee calls in natural habitats is very rare, i.e. databases feature a heavy imbalance between background and target calls. Such imbalances can have negative effects on classifier performances. We employed a state-of-the-art detection approach based on convolutional recurrent neural networks (CRNNs). We extended the detection pipeline through various stages for compensating class imbalance. These included (1) spectrogram denoising, (2) alternative loss functions, and (3) resampling. Our key findings are: (1) spectrogram denoising operations significantly improved performance for both target classes, (2) standard binary cross entropy reached the highest performance, and (3) manipulating relative class imbalance through resampling either decreased or maintained performance depending on the target class. Finally, we reached detection performances of 33 % F 1 for drumming and 5 % F 1 for vocalization, which is a > 7 fold increase compared to previously published results. We conclude that supporting the network to learn decoupling noise conditions from foreground classes is of primary importance for increasing performance.

Pain assessment in clinical settings can be a challenging task for clinic staff and patients. An automatic and contactless pain estimation system via facial expression is a decisive advantage. In the following article we present a method for pain classification based on mimic descriptors. To automatically process facial expressions from video data we applied the Facial Action Coding System (FACS). First, we employed a logistic regression model to predict pain using a publicly available database. Second, we utilized this procedure to predict pain in individual intraprocedural patient data. We reached a mean accuracy of 80.94% for the publicly available data and 73.14% for our individual patient data by classifying data as pain and non-pain.