Accuracy of a markerless acquisition technique for studying speech articulators. In Interspeech 2015

Abstract : The main disadvantages of the existing methods for studying speech articulators (such as electromagnetic and optoelectronic systems) are the high cost and the discomfort to participants or patients. The aim of this work is to introduce a completely markerless low-cost 3D tracking technique in the context of speech articulation, and then compare it with a well-established marker-based one to evaluate the performances. A Kinect-like device was used in conjunction with an existing face tracking algorithm to track lips movements in 3D without markers. The method was tested on two subjects uttering 200 words and 100 sentences. For most of points of the lips the RMSE ranged between 1 and 3 mm. Although the image resolution used in this experiment was low, these results are very promising. Nevertheless, further studies should consider higher video resolutions in order to obtain better results.
Complete list of metadatas

Cited literature [18 references]  Display  Hide  Download

https://hal.inria.fr/hal-01189000
Contributor : Slim Ouni <>
Submitted on : Tuesday, October 6, 2015 - 3:19:58 PM
Last modification on : Wednesday, August 21, 2019 - 9:02:02 PM
Long-term archiving on: Thursday, January 7, 2016 - 10:36:30 AM

File

837_Paper-Bandini.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01189000, version 1

Collections

Citation

Andrea Bandini, Slim Ouni, Piero Cosi, Silvia Orlandi, Claudia Manfredi. Accuracy of a markerless acquisition technique for studying speech articulators. In Interspeech 2015. Interspeech 2015, Sep 2015, Dresden, Germany. ⟨hal-01189000⟩

Share

Metrics

Record views

508

Files downloads

522