We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: 3D-CNN for Facial Micro- and Macro-expression Spotting on Long Video Sequences using Temporal Oriented Reference Frame

Abstract: Facial expression spotting is the preliminary step for micro- and macro-expression analysis. The task of reliably spotting such expressions in video sequences is currently unsolved. The current best systems depend upon optical flow methods to extract regional motion features, before categorisation of that motion into a specific class of facial movement. Optical flow is susceptible to drift error, which introduces a serious problem for motions with long-term dependencies, such as high frame-rate macro-expression. We propose a purely deep learning solution which, rather than track frame differential motion, compares via a convolutional model, each frame with two temporally local reference frames. Reference frames are sampled according to calculated micro- and macro-expression durations. We show that our solution achieves state-of-the-art performance (F1-score of 0.126) in a dataset of high frame-rate (200 fps) long video sequences (SAMM-LV) and is competitive in a low frame-rate (30 fps) dataset (CAS(ME)2). In this paper, we document our deep learning model and parameters, including how we use local contrast normalisation, which we show is critical for optimal results. We surpass a limitation in existing methods, and advance the state of deep learning in the domain of facial expression spotting.
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2105.06340 [cs.CV]
  (or arXiv:2105.06340v2 [cs.CV] for this version)

Submission history

From: Chuin Hong Yap [view email]
[v1] Thu, 13 May 2021 14:55:06 GMT (1524kb,D)
[v2] Thu, 10 Jun 2021 12:39:31 GMT (1558kb,D)

Link back to: arXiv, form interface, contact.