Exploring Edge TPU for deep feed-forward neural networks

Article Agriculture, Multidisciplinary

Deep learning-based cattle behaviour classification using joint time-frequency data representation

Seyedehfaezeh Hosseininoorbin et al.

Summary: This paper introduces a cattle behavior classification method based on time-frequency domain data representation. Experimental results show that this method efficiently reduces model size and computational complexity with only a minor decrease in classification accuracy, achieving high classification performance across different numbers of behavior classes.

COMPUTERS AND ELECTRONICS IN AGRICULTURE (2021)

添加到收藏夹

Article Automation & Control Systems

A study on the use of Edge TPUs for eye fundus image segmentation

Javier Civit-Masot et al.

Summary: The study evaluated the performance of two SBCs for fundus image segmentation, one with hardware acceleration and one without, demonstrating the feasibility of using Machine Learning accelerated SBCs for optic disc and cup segmentation in under 25 ms per image when utilizing Edge TPUs.

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE (2021)

添加到收藏夹

Proceedings Paper Computer Science, Information Systems

Scaling Spectrogram Data Representation for Deep Learning on Edge TPU

Seyedehfaezeh Hosseininoorbin et al.

Summary: This paper explores the use of Google's Edge TPU in automated cattle activity classification, showing that Edge TPU can provide high classification accuracy and energy efficiency. However, it exhibits sensitivity to the chosen neural network model size, which requires appropriate adjustment.

2021 IEEE INTERNATIONAL CONFERENCE ON PERVASIVE COMPUTING AND COMMUNICATIONS WORKSHOPS AND OTHER AFFILIATED EVENTS (PERCOM WORKSHOPS) (2021)

添加到收藏夹

Review Chemistry, Analytical