On the inference speed and video-compression robustness of DeepLabCut

BioRxiv : the Preprint Server for Biology
Alexander Mathis, Richard A. Warren

Abstract

Pose estimation is crucial for many applications in neuroscience, biomechanics, genetics and beyond. We recently presented a highly efficient method for markerless pose estimation based on transfer learning with deep neural networks called DeepLabCut. Current experiments produce vast amounts of video data, which pose challenges for both storage and analysis. Here we improve the inference speed of DeepLabCut by up to tenfold and benchmark these updates on various CPUs and GPUs. In particular, depending on the frame size, poses can be inferred offline at up to 1200 frames per second (FPS). For instance, 278 x 278 images can be processed at 225 FPS on a GTX 1080 Ti graphics card. Furthermore, we show that DeepLabCut is highly robust to standard video compression (ffmpeg). Compression rates of greater than 1,000 only decrease accuracy by about half a pixel (for 640 x 480 frame size). DeepLabCut's speed and robustness to compression can save both time and hardware expenses.

Related Concepts

Biomechanics
Titanium
Psychological Transfer
Geographutoxin II
Size
poly(carbonate urea) urethane
Research Study
Analysis
Trunk of Deep Temporal Nerve
Video Media

Related Feeds

Biomechanics

Biomechanics examines the generation of internal forces within the body and investigates the effects and control of forces that act on or are produced on tissues. Here are the latest discoveries.

BioRxiv & MedRxiv Preprints

BioRxiv and MedRxiv are the preprint servers for biology and health sciences respectively, operated by Cold Spring Harbor Laboratory. Here are the latest preprint articles (which are not peer-reviewed) from BioRxiv and MedRxiv.

Related Papers

BioRxiv : the Preprint Server for Biology
Tanmay NathMackenzie W. Mathis
IEEE Transactions on Information Technology in Biomedicine : a Publication of the IEEE Engineering in Medicine and Biology Society
Gundolf KieferJürgen Weese
© 2021 Meta ULC. All rights reserved