Multimodal Human Action Recognition Based on a Fusion of Dynamic Images Using CNN Descriptors

Edwin Jonathan Escobedo Cardenas, Guillermo Camara Chavez

Producción científica: Capítulo del libro/informe/acta de congresoArticulo (Contribución a conferencia)revisión exhaustiva

6 Citas (Scopus)

Resumen

In this paper, we propose the use of dynamic-images-based approach for action recognition. Specifically, we exploit the multimodal information recorded by a Kinect sensor (RGB-D and skeleton joint data). We combine several ideas from rank pooling and skeleton optical spectra to generate dynamic images to summarize an action sequence into single flow images. We group our dynamic images into five groups: a dynamic color group (DC); a dynamic depth group (DD) and three dynamic skeleton groups (DXY, DYZ, DXZ). As action is composed of different postures along time, we generated N different dynamic images with the main postures for each dynamic group. Next, we applied a pre-trained flow-CNN to extract spatiotemporal features with a max-mean aggregation. The proposed method was evaluated on a public benchmark dataset, the UTD-MHAD, and achieved the state-of-the-art result.

Idioma originalInglés
Título de la publicación alojadaProceedings - 31st Conference on Graphics, Patterns and Images, SIBGRAPI 2018
EditorialInstitute of Electrical and Electronics Engineers Inc.
Páginas95-102
Número de páginas8
ISBN (versión digital)9781538692646
DOI
EstadoPublicada - 15 ene. 2019
Publicado de forma externa
Evento31st Conference on Graphics, Patterns and Images, SIBGRAPI 2018 - Foz do Iguacu, Brasil
Duración: 29 oct. 20181 nov. 2018

Serie de la publicación

NombreProceedings - 31st Conference on Graphics, Patterns and Images, SIBGRAPI 2018

Conferencia

Conferencia31st Conference on Graphics, Patterns and Images, SIBGRAPI 2018
País/TerritorioBrasil
CiudadFoz do Iguacu
Período29/10/181/11/18

Huella

Profundice en los temas de investigación de 'Multimodal Human Action Recognition Based on a Fusion of Dynamic Images Using CNN Descriptors'. En conjunto forman una huella única.

Citar esto