| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 4.7 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Na dobragem de produções audiovisuais só são traduzidas as falas, existindo uma discrepância entre o áudio e o que a personagem diz (movimentos da face, boca e lábios), que resulta em falhas na compreensão da fala e em dobragens pouco realistas. Nos últimos anos, têm sido desenvolvidos métodos de captura e reprodução de movimentos faciais que capturam os movimentos faciais de um ator de dobragem e reproduzem esses movimentos na face de um ator previamente gravado. Este documento contém a análise e avaliação do estado da arte de métodos de captura e de reprodução de movimentos faciais, e a descrição de uma solução de captura e reprodução em tempo-real, utilizando uma câmara normal, desenvolvida para tentar resolver os problemas existentes com as dobragens tradicionais. A solução implementada foi avaliada através de questionários efetuados, demonstrando qualidade ainda inferior às dobragens tradicionais.
In the dubbing of audio-visual productions, only the lines are translated, and there is a discrepancy between the audio and what the character says (movements of the face, mouth, and lips), which results in poor speech comprehension and unrealistic dubs. In recent years, methods of capturing and reproducing facial movements have been developed that capture the facial movements of a dubbing actor and reproduce these movements in the face of a previously recorded actor. This document contains the analysis and evaluation of the state of the art in methods of capture and reproduction of facial movements, and the description of a real-time capture and reproduction solution, using a normal camera, developed to address existing problems with traditional dubbing. The implemented solution was evaluated through questionnaires, showing a quality that is still inferior to traditional dubbing.
In the dubbing of audio-visual productions, only the lines are translated, and there is a discrepancy between the audio and what the character says (movements of the face, mouth, and lips), which results in poor speech comprehension and unrealistic dubs. In recent years, methods of capturing and reproducing facial movements have been developed that capture the facial movements of a dubbing actor and reproduce these movements in the face of a previously recorded actor. This document contains the analysis and evaluation of the state of the art in methods of capture and reproduction of facial movements, and the description of a real-time capture and reproduction solution, using a normal camera, developed to address existing problems with traditional dubbing. The implemented solution was evaluated through questionnaires, showing a quality that is still inferior to traditional dubbing.
Description
Keywords
Captura e reprodução de movimentos faciais Modelos faciais 3D Dobragem Capture and re-enactment of facial movements 3D Morphable Model Dubbing