Toward Scalable Social Alt Text: Conversational Crowdsourcing as a Tool for Refining Vision-to-Language Technology for the Blind
Open accessThe DOI matches its Crossref record; 4 of the 4 references it cites were found in a registry or an index.
Abstract
The access of visually impaired users to imagery in social media is constrained by the availability of suitable alt text. It is unknown how imperfections in emerging tools for automatic caption generation may help or hinder blind users' understanding of social media posts with embedded imagery. In this paper, we study how crowdsourcing can be used both for evaluating the value provided by existing automated approaches and for enabling workflows that provide scalable and useful alt text to blind users. Using real-time crowdsourcing, we designed experiences that varied the depth of interaction of the crowd in assisting visually impaired users at caption interpretation, and measured trade-offs in effectiveness, scalability, and reusability. We show that the shortcomings of existing AI image captioning systems frequently hinder a user's understanding of an image they cannot see to a degree that even clarifying conversations with sighted assistants cannot correct. Our detailed analysis of the set of clarifying conversations collected from our studies led to the design of experiences that can effectively assist users in a scalable way without the need for real-time interaction. They also provide lessons and guidelines that human captioners and the designers of future iterations of AI captioning systems can use to improve labeling of social media imagery for blind users.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Show and tell: A neural image caption generator
Vinyals, O., Toshev, A., Bengio, S., & Erhan, D. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298935 · Cited by 6,318
Deep visual-semantic alignments for generating image descriptions
Karpathy, A., & Fei-Fei, L. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298932 · Cited by 4,902
Understanding Blind People's Experiences with Computer-Generated Captions of Social Media Images
MacLeod, H., Bennett, C. L., Morris, M. R., & Cutrell, E. (2017) · Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3025453.3025814 · Cited by 200 · cited 9×
“These findings are in line with recent findings that BVI users place too much trust in AI captions (MacLeod et al. 2017), and our results suggest that this trust in the initial caption may mislead people to the extent that they are unable to properly frame questions to improve their understanding.”
BackgroundMethodResultInfluentialSelf-citationGlobal Web Accessibility Analysis of National Government Portals and Ministry Web Sites
Goodwin, M., Susar, D., Nietzio, A., Snaprud, M., & Jensen, C. S. (2011) · Journal of Information Technology & Politics · doi:10.1080/19331681.2010.508011 · Cited by 74 · cited 2×
“) that provides computer access to people who are BVI (Kloots 2016); however, while no official numbers on alt text compliance for Twitter are yet available, alt text compliance and quality on the web in general is low (Bigham et al. 2006; Goodwin et al. 2011; Shi 2006), and this trend is likely to be exacerbated by q…”
Background
Cited by
Rahwan, I., Cebrian, M., Obradovich, N., Bongard, J., Bonnefon, J.-F., Breazeal, C., Crandall, J. W., Christakis, N. A., Couzin, I. D., Jackson, M. O., et al. (2019) · Nature · doi:10.1038/s41586-019-1138-y · Cited by 1,364
Self-citationRich Representations of Visual Content for Screen Reader Users
Morris, M. R., Johnson, J., Bennett, C. L., & Cutrell, E. (2018) · Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3173574.3173633 · Cited by 159
“Further, automated captions do not yet include the amount and types of detail desired by many screen reader users [27].”
BackgroundInfluentialSelf-citationStangl, A., Morris, M. R., & Gurari, D. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376404 · Cited by 139
Self-citationBennett, C. L., Gleason, C., Scheuerman, M. K., Bigham, J. P., Guo, A., & To, A. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445498 · Cited by 128
“Researchers have studied several solutions to increase the number of described images from crowdsourcing [93] to generating them automatically [40, 123].”
BackgroundMorris, M. R. (2020) · Communications of the ACM · doi:10.1145/3356727 · Cited by 114
Self-citationCaption Crawler: Enabling Reusable Alternative Text Descriptions using Reverse Image Search
Guinness, D., Cutrell, E., & Morris, M. R. (2018) · Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3173574.3174092 · Cited by 108
Self-citationJung, C., Mehta, S., Kulkarni, A., Zhao, Y., & Kim, Y.-S. (2022) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2021.3114846 · Cited by 91
Toward Automatic Audio Description Generation for Accessible Videos
Wang, Y., Liang, W., Huang, H., Zhang, Y., Li, D., & Yu, L.-F. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445347 · Cited by 87
“proposed a real-time crowd-sourcing experience which brought human in the loop to compensate for the shortcomings of existing AI algorithms [46].”
BackgroundStangl, A., Verma, N., Fleischmann, K. R., Morris, M. R., & Gurari, D. (2021) · Proceedings of the 23rd International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3441852.3471233 · Cited by 81
Self-citationGleason, C., Pavel, A., Liu, X., Carrington, P., Chilton, L. B., & Bigham, J. P. (2019) · Proceedings of the 21st International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3308561.3353792 · Cited by 69
BrowseWithMe: An Online Clothes Shopping Assistant for People with Visual Impairments
Stangl, A. J., Kothari, E., Jain, S. D., Yeh, T., Grauman, K., & Gurari, D. (2018) · Proceedings of the 20th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3234695.3236337 · Cited by 66
SciCap: Generating Captions for Scientific Figures
Hsu, T.-Y., Giles, C. L., & Huang, T.-H. (2021) · Findings of the Association for Computational Linguistics: EMNLP 2021 · doi:10.18653/v1/2021.findings-emnlp.277 · Cited by 59
Measuring Representational Harms in Image Captioning
Wang, A., Barocas, S., Laird, K., & Wallach, H. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533099 · Cited by 40
“3Although it falls outside the scope of this paper, we note that there is a worrisome disconnect between technical research on image captioning, which often uses accessibility as a motivation, and usability research focused on the real-world value of image captioning [49, 50, 54, 63, 64].”
BackgroundComputer Vision and Conflicting Values: Describing People with Automated Alt Text
Hanley, M., Barocas, S., Levy, K., Azenkot, S., & Nissenbaum, H. (2021) · Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3461702.3462620 · Cited by 38
Chintalapati, S. S., Bragg, J., & Wang, L. L. (2022) · Proceedings of the 24th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3517428.3544796 · Cited by 31
Nair, V., Zhu, H. '., & Smith, B. A. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581302 · Cited by 29
“Prior research, however, has shown that alt text authored by humans may sufer from quality issues [11, 33, 59], and that autogenerated descriptions may be error-prone and misleading [31, 49].”
BackgroundToward supporting quality alt text in computing publications
Williams, C., de Greef, L., Harris, E., Findlater, L., Pavel, A., & Bennett, C. (2022) · Proceedings of the 19th International Web for All Conference · doi:10.1145/3493612.3520449 · Cited by 26
Blind Users Accessing Their Training Images in Teachable Object Recognizers
Hong, J., Gandhi, J., Mensah, E. E., Zeraati, F. Z., Jarjue, E., Lee, K., & Kacorri, H. (2022) · Proceedings of the 24th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3517428.3544824 · Cited by 22
“Thus, it is not a surprise to find that recognition errors affect blind users’ experience [51]; sometimes, to a degree where it can not be corrected even by human clarification [61].”
Background(2019) · doi:10.1145/3308561.3353808 · Cited by 14
Understanding How Blind Users Handle Object Recognition Errors: Strategies and Challenges
Hong, J., & Kacorri, H. (2024) · The 26th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663548.3675635 · Cited by 14
Crowdsourcing Thumbnail Captions: Data Collection and Validation
Aguirre, C., Cao, S., Mahmood, A., & Huang, C.-M. (2023) · ACM Transactions on Interactive Intelligent Systems · doi:10.1145/3589346 · Cited by 6
Metadata
- Authors
- Salisbury, E., Kamar, E., & Morris, M.
- Venue
- Proceedings of the AAAI Conference on Human Computation and Crowdsourcing
- Published
- 2017-09-21
- Volume
- 5
- Pages
- 147-156
- ISSN
- 2769-1330 (ISSN-L)
- 2769-1349
- OpenAlex
- W2748163013
- Cited by
- 83
- References
- 4
- FWCI
- 10.10
- Type
- conference paper
- Language
- en
- Open access
- diamond
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.