Analog Gated Recurrent Neural Network for Detecting Chewing Events
Author ORCID Identifier
Kofi Odame: https://orcid.org/0000-0002-8519-2392
Maria Nyamukuru: https://orcid.org/0000-0002-7748-5013
Mohsen Shahghasemi: https://orcid.org/0000-0003-0638-516X
David Kotz: https://orcid.org/0000-0001-7411-2783
IEEE Transactions on Biomedical Circuits and Systems
Thayer School of Engineering
Department of Computer Science
We present a novel gated recurrent neural network to detect when a person is chewing on food. We implemented the neural network as a custom analog integrated circuit in a 0.18 μm CMOS technology. The neural network was trained on 6.4 hours of data collected from a contact microphone that was mounted on volunteers’ mastoid bones. When tested on 1.6 hours of previously-unseen data, the analog neural network identified chewing events at a 24-second time resolution. It achieved a recall of 91% and an F1-score of 94% while consuming 1.1 μW of power. A system for detecting whole eating episodes— like meals and snacks— that is based on the novel analog neural network consumes an estimated 18.8 μW of power.
Kofi Odame, Maria Nyamukuru, Mohsen Shahghasemi, Shengjie Bi, and David Kotz. Analog Gated Recurrent Neural Network for Detecting Chewing Events. IEEE Transactions on Biomedical Circuits and Systems, volume 16, number 6, pages 1106–1115. IEEE, December 2022. doi:10.1109/TBCAS.2022.3218889.
Dartmouth Digital Commons Citation
Odame, Kofi; Nyamukuru, Maria; Shahghasemi, Mohsen; Bi, Shengjie; and Kotz, David, "Analog Gated Recurrent Neural Network for Detecting Chewing Events" (2022). Dartmouth Scholarship. 4310.