Lessons Learned from Crowd Accessibility Services - Human-Computer Interaction – INTERACT 2013 Access content directly
Conference Papers Year : 2013

Lessons Learned from Crowd Accessibility Services

Hironobu Takagi
  • Function : Author
  • PersonId : 1004976
Susumu Harada
  • Function : Author
  • PersonId : 1004977
Daisuke Sato
  • Function : Author
  • PersonId : 1004978
Chieko Asakawa
  • Function : Author
  • PersonId : 1004979

Abstract

Crowd accessibility services for people with disabilities, driven by crowd-sourcing methods, are gaining traction as a viable means of realizing innovative services by leveraging both human and machine intelligence. As the approach matures, researchers and practitioners are seeking to build various types of services. However, many of them encounter similar challenges, such as variations in quality and sustaining contributor participation for durable services. There are growing needs to share tangible knowledge about the best practices to help build and maintain successful services. Towards this end, we are sharing our experiences with crowd accessibility services that we have deployed and studied. Initially, we developed a method to analyze the dynamics of contributor participation. We then analyzed the actual data from three service deployments spanning several years. The service types included Web accessibility improvement, text digitization, and video captioning. We then summarize the lessons learned and future research directions for sustainable services.
Fichier principal
Vignette du fichier
978-3-642-40483-2_42_Chapter.pdf (405.36 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01497465 , version 1 (28-03-2017)

Licence

Attribution

Identifiers

Cite

Hironobu Takagi, Susumu Harada, Daisuke Sato, Chieko Asakawa. Lessons Learned from Crowd Accessibility Services. 14th International Conference on Human-Computer Interaction (INTERACT), Sep 2013, Cape Town, South Africa. pp.587-604, ⟨10.1007/978-3-642-40483-2_42⟩. ⟨hal-01497465⟩
99 View
63 Download

Altmetric

Share

Gmail Facebook X LinkedIn More