Evaluation of System of Automated Audio description Script Generation: An Analysis from the User Perspective

• Rafael Cordeiro UFPB
• Tiago Araújo UFPB
• Yuska Aguiar UFPB

DOI: https://doi.org/10.5753/ihc.2019.8413

Abstract

Audio Description (known as AD) is a narrative audio track that helps visually impaired people to understand audiovisual content. The professional responsible for generating AD is the describer. He/She has the task of turning the visual information of videos or presentations into audio for viewers. This transformation process is time-consuming and costly. In this context, the development of system of audio description script generation is an attractive solution to support describer and favorite the reduction of time and cost on this translation process. This research aims to evaluate the audio description generated automatically, for short-films, by two computational models system and a human description. The proposed methodologies focus on the quality of AD generated from the perspective of the visually impaired user. The data collected was performed by a semi-structured interview. The accessibility level of audio descriptions was associated with the participants' comprehension about short film contents. The results show the automatic script-based AD model. The most recent results in human AD, while the object-based model contributed minimally to comprehension, with inferior results in the scenario without audio description.

Keywords: Audio Description, Accessibility, User experience.

References

ACO Lima, MFQ Vieira, R.S Ferreira, YPC Aguiar, MP Bastos and SLM Lopes Júnior. Evaluating system accessibility using an experimental protocol based on usability. International Conferences on Interfaces and Human Computer Interaction 2018, 2018. p. 85-92.

E. P. C Franco and M. C. C. C SILVA. Audiodescrição: breve passeio histórico. Audiodescrição: transformando imagens em palavras. São Paulo: Secretaria dos Direitos da Pessoa com Deficiência do Estado de São Paulo, p.23-42, 2010.

F. J. de Lima. Áudio-descrição: orientações para uma prática sem barreiras atitudinais. Revista Brasileira de Tradução Visual, v. 2, n. 2, 2010.

IBGE. Censo demográfico 2010. IBGE: Instituto Brasileiro de Geografia e Estatística,2010.

J. Lakritz and A. Salway, The Semi-Automatic Generation of Audio Description from Screenplays, Dept. Of Computing Technical Report CS-06-05, 2002, University of Surrey. http://www.bbrel.co.uk/pdfs/CS-06-05.pdf

R. M. Correia. Uma Revisão Sistemática da Literatura sobre Técnicas de Geração Automática e Semi-Automática. João Pessoa: UFPB, 2014.

V. C. D. S. P. Brasil, (2000). Lei nº 10.098, de 19 de dezembro de 2000. Diário Oficial da União, https://www2.camara.leg.br/legin/fed/lei/2000/lei-10098-19-dezembro-2000-377651-publicacaooriginal-1-pl.html

VP Campos. Um sistema de geração automática de roteiros de audiodescrição. 2015. 89 f. Dissertação (Mestrado em Informática) - Universidade Federal da Paraíba, João Pessoa, 2015.

VP Campos, LM Gonçalves and TMU De Araújo. Applying audio description for context understanding of surveillance videos by people with visual impairments. In: 2017 14th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). IEEE, 2017. p. 1-5.

W.S. Pereira et al. Usuários com Síndrome de Down: Como realizar Testes de Usabilidade?. RENOTE, v. 15, n. 2. https://doi.org/10.22456/1679-1916.79274

Y. P. C Aguiar. (2012). Protocolo Experimental para Observação da Interação um Arcabouço para o Estudo do Comportamento Humano. Campina Grande: UFCG.