There is a growing interest in low power highly efficient wearable devices for automatic dietary monitoring (ADM) [1]. The success of deep neural networks in audio event classification problems makes them ideal for this task. Deep neural networks are, however, not only computationally intensive and energy inefficient but also require a large amount of memory.To address these challenges, we propose a shallow gated recurrent unit (GRU) architecture suitable for resource-constrained applications. This paper describes the implementation of the Tiny Eats GRU, a shallow GRU neural network, on a low power microcontroller, Arm Cortex M0+, to classify eating episodes. Tiny Eats GRU is a hybrid of the traditional GRU [2] and eGRU [3] which makes it small and fast enough to fit on the Arm Cortex M0+ with comparable accuracy to the traditional GRU. The Tiny Eats GRU utilizes only 4% of the Arm Cortex M0+ memory and identifies eating or non-eating episodes with 6 ms latency and accuracy of 95.15%.

No abstract

At-home monitoring of lung health enables the early detection and treatment of respiratory diseases like asthma and chronic obstructive pulmonary disease (COPD). To allow for discreet continuous monitoring, various approaches have been proposed to estimate the respiratory rate from an electrocardiogram (ECG) signal. Unfortunately, respiratory rate can only provide a non-specific, incomplete picture of lung health. This paper introduces an algorithm to extract more respiratory information from the ECG signal: in addition to respiratory rate, the algorithm also derives the fractional inspiratory time(FIT), which is a direct measure of airway obstruction. The algorithm is based on a gated recurrent neural network that infers vital respiratory information from a two-lead ECG signal. The network is trained and tested on different test subjects and reports up to 0.099, and 0.243 normalized root mean squared error in the computation of FIT and respiration rate, respectively.

No abstract

No abstract