The DOI matches its Crossref record; 44 of the 45 references it cites were found in a registry or an index.
VizWiz Grand Challenge: Answering Visual Questions from Blind People
Open accessAbstract
The study of algorithms to automatically answer visual questions currently is motivated by visual question answering (VQA) datasets constructed in artificial VQA settings. We propose VizWiz, the first goal-oriented VQA dataset arising from a natural VQA setting. VizWiz consists of over 31,000 visual questions originating from blind people who each took a picture using a mobile phone and recorded a spoken question about it, together with 10 crowdsourced answers per visual question. VizWiz differs from the many existing VQA datasets because (1) images are captured by blind photographers and so are often poor quality, (2) questions are spoken and so are more conversational, and (3) often visual questions cannot be answered. Evaluation of modern algorithms for answering visual questions and deciding if a visual question is answerable reveals that VizWiz is a challenging dataset. We introduce this dataset to encourage a larger community to develop more generalized algorithms that can assist blind people.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Prior work reported over 40% disagreement between crowdworkers and in-house experts on whether a photo contained privacy or quality issues [36]”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [36]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Although academic scholars and industry corporations have emphasized the pressing need for more disability-centered datasets [20, 38, 55, 79, 97, 111, 124]”
From "It's trained by non-disabled people": Evaluating How Image Quality Affects Product Captioning with Vision-Language Models (2026) · cites this paper as [55]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“While prior efforts of disabilityfirst datasets often focus on the data collection process [36, 47, 63, 68, 88, 98], our study represents an important step in centering disabled people and their expertise throughout the AI development pipeline.”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [36]
The cited source says what the claim attributes to it.
In the citing paper
“From analyzing VizWiz images, Gurari et al. [55] found that blind users often struggle to take high-quality photographs, and many visual questions go unanswered because images fail to capture the relevant objects [55].”
From "It's trained by non-disabled people": Evaluating How Image Quality Affects Product Captioning with Vision-Language Models (2026) · cites this paper as [55]
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
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 2×
“Following prior work [14], we first computed the average image from all images in VizWiz.”
BackgroundMethodMicrosoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., & Zitnick, C. L. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_48 · Cited by 45,460 · cited 4×
“0: Real [7] MSCOCO [25] Crowd workers (AMT) Typed Visual Madlibs [43] MSCOCO [25] Automatically generated (templates) ——– FM-IQA [16] MSCOCO [25] Crowd workers (Baidu) Typed KB-VQA [40] MSCOCO [25] In-house participants Typed COCO-QA [32] MSCOCO [25] Automatically generated (captions) ——– VQA v2.”
BackgroundMethodInfluentialImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
BLEU: a method for automatic evaluation of machine translation
Papineni, K., Roukos, S., Ward, T., & Zhu, W.-J. (2001) · Proceedings of the 40th Annual Meeting on Association for Computational Linguistics - ACL '02 · doi:10.3115/1073083.1073135 · Cited by 21,344 · cited 2×
“We employ five metrics commonly used for evaluating VQA and image description algorithms: Accuracy [7], CIDEr [38], BLEU4 [29], and ME-”
MethodIndoor segmentation and support inference from rgbd images
Silberman, N., Hoiem, D., Kohli, P., & Fergus, R. (2012) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-33715-4_54 · Cited by 5,760 · cited 1×
“DAQUAR [28] NYU Depth V2 [34] Automatically generated (templates) ——– VQA v1.”
MethodVisual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations
Krishna, R., Zhu, Y., Groth, O., Johnson, J., Hata, K., Kravitz, J., Chen, S., Kalantidis, Y., Li, L.-J., Shamma, D. A., et al. (2017) · International Journal of Computer Vision · doi:10.1007/s11263-016-0981-7 · Cited by 5,107
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
CIDEr: Consensus-based image description evaluation
Vedantam, R., Zitnick, C. L., & Parikh, D. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7299087 · Cited by 4,845 · cited 2×
“We employ five metrics commonly used for evaluating VQA and image description algorithms: Accuracy [7], CIDEr [38], BLEU4 [29], and ME-”
MethodVQA: Visual Question Answering
Antol, S., Agrawal, A., Lu, J., Mitchell, M., Batra, D., Zitnick, C. L., & Parikh, D. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.279 · Cited by 4,487
Bleu: a method for automatic evaluation of machine translation
Papineni, K. (2024) · doi:10.57702/ljteqmrb · Cited by 3,668
Flagged first, then by citations
1 of the 45 references couldn’t be matched to a known paper.
Research works printed but not linked: 3
- mahendru, 2017, The promise of premise Harnessing question premises in visual question answering
- anderson, 2017, Bottom-up and top-down attention for image captioning and visual question answering
- 0, Be My Eyes
Cited by
Sorted by citations
Items 301–400 of 629
Lemmer, S. J., Guo, A., & Corso, J. J. (2023) · ACM International Conference on Intelligent User Interfaces (IUI) · doi:10.1145/3581641.3584092 · Cited by 9
“Other settings have similar limitations: vision-and-dialog navigation [78] also has significant latency after inference, while visual question answering for the visually impaired [31] does not have a simplemethod for confirming the model’s output.”
BackgroundInfluentialSelf-citationBenchmarking Out-of-Distribution Detection in Visual Question Answering
Shi, X., & Lee, S. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00540 · Cited by 6
“AUCROC Upper Bounds for VizWiz and GQA.”
BackgroundMethodInfluentialMMInstruct: a high-quality multi-modal instruction tuning dataset with extensive diversity
Liu, Y., Cao, Y. Q., Gao, Z., Wang, W., Chen, Z., Wang, W., Tian, H., Lu, L., Zhu, X., Lu, T., et al. (2024) · Science China Information Sciences · doi:10.1007/s11432-024-4187-3 · Cited by 34
“These results clearly demonstrate that MMInstruct benefits visual instruction tuning.”
Improving Generalization in Visual Reasoning via Self-ensemble
Nguyen, T.-H., Tran, Q.-K., & Quang-Hoang, A.-T. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-91672-4_18 · Cited by 12
TA-Student VQA: Multi-Agents Training by Self-Questioning
Xiong, P., & Wu, Y. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01008 · Cited by 7
“In recent years, Visual Question Answering ( VQA ) has garnered significant attention [45, 16, 35, 51], as it relates to multidisciplinary research such as natural language understanding [58], visual information retrieval [19, 62], and multi-modal reasoning [4, 39].”
BackgroundVisual World to an Audible Experience: Visual Assistance for the Blind And Visually Impaired
Mohith, S. S., Vijay, S., Sanjana, V., & Krupa, N. (2020) · IEEE India Conference · doi:10.1109/indicon49873.2020.9342481 · Cited by 6
“Hence we would like to train our models on VizWiz dataset [23] as the images were taken by the blind people and the questions were also asked by the blind people.”
MethodMulti-Tier Attention Network using Term-weighted Question Features for Visual Question Answering
Manmadhan, S., & Kovoor, B. C. (2021) · Image and Vision Computing · doi:10.1016/j.imavis.2021.104291 · Cited by 6
MSD: Saliency-aware Knowledge Distillation for Multimodal Understanding
Jin, W., Sanjabi, M., Nie, S., Tan, L., Ren, X., & Firooz, H. (2021) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2021.findings-emnlp.302 · Cited by 10
“For example, visual question answering (VQA) requires answering questions about an image (Antol et al., 2015; Goyal et al., 2017; Gurari et al., 2018; Singh et al., 2019) and models should incorporate information from the text and image sources to answer the questions.”
BackgroundHow to Practice VQA on a Resource-limited Target Domain
Zhang, M., Hwa, R., & Kovashka, A. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00443 · Cited by 6
“To simplify the cross-dataset evaluation, we follow traditions to formulate the task as a multi-way classification, 1For example, popular medical VQA dataset ImageCLEF-2019 [3] has 16K Q&A pairs on 4K radiology images; VizWiz [13] contains 31K images/questions for assisting the visually impaired; the advertisement und…”
BackgroundMethodInfluentialWhy Did the Chicken Cross the Road? Rephrasing and Analyzing Ambiguous Questions in VQA
Stengel-Eskin, E., Guallar-Blasco, J., Zhou, Y. X., & Durme, B. V. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.569 · Cited by 13
Sentence Attention Blocks for Answer Grounding
Khoshsirat, S., & Kambhamettu, C. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00559 · Cited by 6
“The VizWiz-VQA-Grounding dataset [3] is based on the VizWiz-VQA dataset [10], and the images and questions come from visually impaired people who shared them to request visual assistance in their day-to-day lives.”
MethodVisually Grounded Language Learning: A Review of Language Games, Datasets, Tasks, and Models
Suglia, A., Konstas, I., & Lemon, O. (2024) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.15185 · Cited by 10
YTCommentQA: Video Question Answerability in Instructional Videos
Yang, S., Park, S., Jang, Y., & Lee, M. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i17.29906 · Cited by 11
Multimodal Visual Question Answering Using VizWiz Data; A Visual Assistant for the Blind
Sreedha, B., & Nair, P. (2022) · Lecture notes in electrical engineering · doi:10.1007/978-981-19-1742-4_30 · Cited by 6
Gradations of Error Severity in Automatic Image Descriptions
van Miltenburg, E., Lu, W.-T., Krahmer, E., Albert Gatt, A., Chen, G., Li, L., & van Deemter, K. (2020) · International Conference on Natural Language Generation · doi:10.18653/v1/2020.inlg-1.45 · Cited by 6
“As an example, consider the setting of the VizWiz challenge (Gurari et al., 2018), which consists of questions asked by visually impaired people seeking help from online users, based on photographs usually taken using phones.”
MethodLemmer, S. J., Song, J. Y., & Corso, J. J. (2021) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445181 · Cited by 6
“it is assumed that the seed—the question—always requests information available in the primary input—the text document or image, keypoint-conditioned viewpoint estimation [73], where it is assumed that a given click always indicates the corresponding keypoint on the image, and others [6, 24, 26, 30, 32, 55, 60, 70, 72].”
BackgroundAgarwal, P., Betancourt, A., Panagiotou, V., & Díaz-Rodríguez, N. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.11743 · Cited by 6
“Guiding a visually challenged person (Gurari et al., 2018) is a highly sensitive task and a small error can lead to catastrophic accidents.”
BackgroundORBIT: A Real-World Few-Shot Dataset for Teachable Object RecognitionPreprint
Massiceti, D., Zintgraf, L., Bronskill, J. F., Theodorou, L., Harris, M. T., Cutrell, E. B., Morrison, C. P., Hofmann, K., & Stumpf, S. (2021) · arXiv (Cornell University) · doi:10.25383/city.14294597 · Cited by 6
BIV-Priv-Seg: Locating Private Content in Images Taken by People With Visual Impairments
Tseng, Y., Sharma, T., Zhang, L., Stangl, A. J., Findlater, L., Wang, Y., & Gurari, D. (2025) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61041.2025.00052 · Cited by 12
Self-citationTang, Y., Fang, Y., Wang, T., Sun, L., & Chen, L. (2025) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3746059.3747597 · Cited by 7
“The two datasets were then combined, resulting in a final dataset of 619 hallucination cases. includes over 31,000 real-life VQA cases sourced from BLV people worldwide [33].”
InfluentialYu, H., Magay, A., Huang, H. M., Yuan, S., Wen, C., & Fang, Y. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801606 · Cited by 8
“Experimental Setting: In this section, we conduct experiments to test the effectiveness of each component within our system on the task of Visual Question Answering [26] of VizWiz dataset.”
VCF: An effective Vision-Centric Framework for Visual Question Answering
Wang, F., Peng, L., Cao, S., Yang, Z., Zhang, R., & An, G. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.129536 · Cited by 5
Anomaly Detection for People with Visual Impairments Using an Egocentric 360-Degree Camera
Song, I., Lee, S., Joo, M., & Lee, J. (2025) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61041.2025.00280 · Cited by 5
“As a result, there has been a significant increase in developing a visual aid system to create new “eyes” for visually impaired individuals using wearable devices and Artificial Intelligence (AI) technologies [12, 26].”
Counterfactual Dual-Bias VQA: A Multimodality Debias Learning for Robust Visual Question Answering
Wang, B., Ju, X., Gao, J., Li, X., Hu, Y., & Yin, B. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3562085 · Cited by 7
“Moreover, in the realm of AI-assisted for the visually impaired, VQA can help limited vision patients understand their surrounding environment [9].”
Ma, X., Xie, H., & Qin, S. J. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103419 · Cited by 7
Testoni, A., Greco, C., & Bernardi, R. (2022) · Frontiers in Big Data · doi:10.3389/fdata.2021.736709 · Cited by 5
“Clearly, they are an important tool for instance as personal assistants of visually impaired people (Gurari et al., 2018), but again their impressive achievements would be vanished if they fail to distinguish negative and affirmative information.”
BackgroundOpportunities and Challenges in Development of Support System for Visually Impaired: A Survey
Vijetha, U., & Geetha, V. (2023) · International Conference on Cloud Computing, Data Science & Engineering (Confluence) · doi:10.1109/confluence56041.2023.10048861 · Cited by 5
“The queries are often conversational-type statements, may or may not have any questions in them, or may have several questions in one conversation [3].”
BackgroundMethodInfluentialAssistSR: Task-oriented Video Segment Retrieval for Personal AI Assistant
Lei, W., Gao, D., Wang, Y., Mao, D., Liang, Z., Ran, L., & Shou, M. Z. (2022) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2022.findings-emnlp.24 · Cited by 5
“…each frame, we used Resnet-50 (He et al., 2016) with weights from grid-feat (Jiang et al., 2020), which is trained on Visual Genome (Krishna et al., 2016) for object detection and attribute classification and produces effective features for image VQA tasks (An-tol et al., 2015; Gurari et al., 2018).”
MethodMulti-Domain Lifelong Visual Question Answering via Self-Critical Distillation
Lao, M., Pu, N., Liu, Y., Zhong, Z., Bakker, E. M., Sebe, N., & Lew, M. S. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612121 · Cited by 5
“…to these settings that focus on inner-domain incremental VQA within a single dataset, we argue that the domain-incremental setting is more practical yet under-explored Figure 1: (a): MDL-VQA involves five tasks in different domains: CLEVR [27], GQA [24], Vizwiz [20], VQA-ab [18], AQUA [15].”
BackgroundInfluentialSCAP: enhancing image captioning through lightweight feature sifting and hierarchical decoding
Zhang, Y., Tong, J., & Liu, H. (2025) · The Visual Computer · doi:10.1007/s00371-025-03824-w · Cited by 5
“These limitations underscore the necessity for models that can efficiently merge visual and textual information in a lightweight manner to achieve optimal IC performance [16].”
Pham, H. Q., Nguyen, T. K.-B., Nguyen, Q. V., Tran, D. Q., Nguyen, N. H., Nguyen, K. V., & Nguyen, N. L.-T. (2025) · Multimedia Systems · doi:10.1007/s00530-025-01696-7 · Cited by 10
“We hired ten native Vietnamese speakers; each person was required to annotate at least 30 questions, divided equally [18] 9285 – – VizWiz [22] 32,842 265,420 265,420 OK-VQA [23] 14,031 14,055 14,055 GQA [24] 113,018 22,669,678 22,669,678 Visual Genome [25] 108,077 into specific fields.”
CoMM: A Coherent Interleaved Image-Text Dataset for Multimodal Understanding and Generation
Chen, W., Lin, L., Yang, Y., Wen, B., Yang, F., Gao, T., Wu, Y., & Chen, L. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00756 · Cited by 5
“This superior performance underlines the strength of our dataset in enabling models to better learn from context and apply learned knowledge to unseen instances.”
Identification of Important Images for Understanding Web Pages
Zhong, Y. C., Matsubara, M., & Morishima, A. (2018) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata.2018.8621911 · Cited by 5
Duan, M., Kumar, A., Saugstad, M., Zeng, A., Savin, I., & Froehlich, J. E. (2021) · International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3441852.3476542 · Cited by 5
“There is growing recognition of the importance in collecting and publishing open datasets in the accessibility community [6,8,9].”
BackgroundWu, X., Zheng, D., Wang, R., Sun, J., Hu, M., Feng, F., Wang, X., Jiang, H., & Yang, F. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548172 · Cited by 5
“The success of VQA on natural images has greatly prompted research in the cross-modal field and drew surging attention [1, 2, 8, 11, 12, 14, 21, 25, 28, 34].”
BackgroundPseudo Content Hallucination for Unpaired Image Captioning
Ben, H., Wang, S., Wang, M., & Hong, R. (2024) · International Conference on Multimedia Retrieval · doi:10.1145/3652583.3658080 · Cited by 5
“It is widely used in many applications, such as image search engines [9, 26, 52], assisting technologies [19, 21, 22, 53, 56], automatic image annotation [10, 14], etc.”
MethodEstimation Of Visual Contents Based On Question Answering From Human Brain Activity
Takada, S., Togo, R., Ogawa, T., & Haseyama, M. (2020) · Proceedings - International Conference on Image Processing · doi:10.1109/icip40778.2020.9191015 · Cited by 5
“In addition, VQA has the possibility of extensive applications in practice, such as assisting visual disability users [6].”
BackgroundBreaking Shortcuts by Masking for Robust Visual Reasoning
Ye, K., Zhang, M., & Kovashka, A. (2021) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv48630.2021.00356 · Cited by 5
“Visual reasoning is an important family of problems including visual question answering (VQA) [5, 9, 12, 41] and visual commonsense reasoning (VCR) [56].”
BackgroundNeural Networks for Detecting Irrelevant Questions During Visual Question Answering
Li, M., Weber, C., & Wermter, S. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-61616-8_63 · Cited by 5
“This task has received significant interest from researchers because it not only can be utilized to examine the development of multimodal and crossmodal technologies [6], but also has great potentials in real-world application scenarios [8].”
BackgroundSelf-Supervised VQA: Answering Visual Questions using Images and Captions.Preprint
Banerjee, P., Gokhale, T., Yang, Y., & Baral, C. R. (2020) · arXiv (Cornell University) · Cited by 4
Alharbi, R., Cheong, A. D., Herskovitz, J., Brewer, R. N., & Schoenebeck, S. (2025) · Proceedings of the 27th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663547.3746376 · Cited by 3
“However, this approach would not necessarily solve the inaccuracy issues discussed in 4.1.1, especially given how object recognition techniques performs poorly on images taken by blind people [53, 77] and for non-Western objects [7, 37].”
InfluentialTang, X., Li, J., & Wu, S. (2026) · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790405 · Cited by 4
“Prior work reported over 40% disagreement between crowdworkers and in-house experts on whether a photo contained privacy or quality issues [36], and only 9% unanimous agreement among annotators on the answer to associated visual questions [35].”
InfluentialSalous, M., Lange, D., von Reeken, T., Wolters, M. K., Heuten, W., Boll, S., & Abdenebaoui, L. (2025) · IEEE Access · doi:10.1109/access.2025.3605490 · Cited by 4
“The associate editor coordinating the review of this manuscript and approving it for publication was Alex James . that are sourced through crowdwork platforms like Amazon Mechanical Turk (e.g. [ 1], [2], [3], [4], [5]).”
Behind the Magic, MERLIM: Multi-Modal Evaluation Benchmark for Large Image-Language Models
Villa, A., Alcázar, J. L., Soto, Á., & Ghanem, B. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00054 · Cited by 5
“In contrast, MERLIM focuses on querying and attributing the response of the IT-LVLM to a specific set of object instances in the image by means of small image alterations and direct verification of the same query on the modified visual data.”
Sharma, T., Tseng, Y., Zhang, L., Ide, A., Mack, K., Findlater, L., Gurari, D., & Wang, Y. (2025) · International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3663547.3746335 · Cited by 6
Self-citationZhang, X., Zhu, Y., Gu, C., Cao, J., Cheng, H., & Wu, K. (2025) · Information Processing & Management · doi:10.1016/j.ipm.2025.104431 · Cited by 4
GeoVisA11y: An AI-based Geovisualization Question-Answering System for Screen-Reader Users
Li, C., Pang, R. Y., Chheda-Kothary, A., Sharif, A., Assalif, H., Heer, J., & Froehlich, J. E. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790334 · Cited by 5
“Research in both image and visualization QA has rejected this assumption, demonstrating that BLV users pose significantly different questions with distinct phrasing patterns [18, 27, 33].”
SMoLoRa: Exploring and Defying Dual Catastrophic Forgetting in Continual Visual Instruction Tuning
Wang, Z., Che, C., Wang, Q., Li, Y., Shi, Z., & Wang, M. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00024 · Cited by 22
VQAsk: a multimodal Android GPT-based application to help blind users visualize pictures
Marsico, M. D., Giacanelli, C., Manganaro, C. G., Palma, A., & Santoro, D. (2024) · International Conference on Advanced Visual Interfaces (AVI) · doi:10.1145/3656650.3656677 · Cited by 4
“To evaluate VQA applications designed to this aim, [6] introduces the VisWiz dataset with 31,000 visual questions from blind people, who took a photo with their mobile phones and recorded a spoken question about it.”
MethodKhan, M. J., Siddiqui, A. M., Khan, H. S., & Khan, M. J. (2025) · International Journal of Data Science and Analytics · doi:10.1007/s41060-025-00827-7 · Cited by 4
ChitroJera: A Regionally Relevant Visual Question Answering Dataset for Bangla
Barua, D. D., Sourove, M. S. U. R., Fahim, M., Haider, F., Shifat, F. T., Adib, M. T. R., Uddin, A. B., Ishmam, M. F., & Alam, M. F. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-06078-5_27 · Cited by 4
A Multilingual Approach to Scene Text Visual Question Answering
Pujolràs, J. B. I., Gómez, L., & Karatzas, D. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-06555-2_5 · Cited by 4
VizWiz grand challenge workshop at CVPR 2022
Massiceti, D., Anjum, S., & Gurari, D. (2022) · ACM SIGACCESS Accessibility and Computing · doi:10.1145/3560232.3560233 · Cited by 4
Self-citationKim, G., Lee, H., Kim, D., Jung, H., Park, S., Kim, Y., Yun, S., Kil, T., Lee, B., & Park, S. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.735 · Cited by 12
UniCode: Learning a Unified Codebook for Multimodal Large Language Models
Zheng, S., Zhou, B., Feng, Y., Wang, Y., & Lu, Z. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73242-3_24 · Cited by 15
Li, Q., & Wu, S. (2024) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3687014 · Cited by 8
“Similarly, Zhao et al. demonstrated that vision and language models could amplify existing gender biases present in training data if not appropriately constrained [38].”
BackgroundInfluentialSelf-citationModality-experts coordinated adaptation for large multimodal models
Zhang, Y., Ji, Z., Pang, Y., Han, J., & Li, X. (2024) · Science China Information Sciences · doi:10.1007/s11432-024-4234-4 · Cited by 4
“On the VisWiz [56], MMBench [57], and the adversarial split of POPE [55] benchmarks, our ModeX also secures the second-best results.”
Continual Instruction Tuning for Large Multimodal Models
He, J., Guo, H., Zhu, K., Tang, M., & Wang, J. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3671616 · Cited by 59
MMFuser: Multimodal Multi-layer Feature Fuser for Fine-Grained Vision-Language Understanding
Cao, Y., Huang, Y., Zhu, W., Liu, Y., Chen, Z., Shi, G., Fa, Y., Yang, Y., Mei, S., & Lü, T. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-31666-0_27 · Cited by 16
Rephrasing Visual Questions by Specifying the Entropy of the Answer Distribution
Terao, K., Tamaki, T., Raytchev, B., Kaneda, K., & Satoh, S. (2020) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2020edp7089 · Cited by 4
“These tasks have been attracting a lot of attention in recent years because it has a large potential to impact many applications such as smart support for the visually impaired [3], providing instructions to autonomous robots [4], and intelligent interaction between humans and machines [5].”
BackgroundExploring Weaknesses of VQA Models through Attribution Driven Insights
Halbe, S. (2020) · doi:10.18653/v1/2020.challengehml-1.9 · Cited by 4
“Further we verify and test these attacks on the following models : (1) Pythia (Singh et al., 2019) (the VizWiz 2018 challenge winner) pretrained on VQA v2 and transferred to VizWiz (train split) and (2) Q+I+A model (which was used to benchmark on VizWiz) trained from scratch on VizWiz (train split).”
BackgroundMethodInfluentialLiterature Review of Computer Tools for the Visually Impaired: A Focus on Search Engines
Meyer, G., Wassyng, A., Lawford, M. S., Sabri, K., & Shirani, S. (2022) · CRC Press eBooks · doi:10.1201/9781003120902-10 · Cited by 4
“Only a fraction of retrieved documents are relevant (p1) - Figure 1: Shows PCA for information access in web SE (p2) Two types of searches: Novice Searches : “users do not have some prior knowledge to search specific information” (p2) Expert Searches : “users usually have this knowledge to search related information b…”
BackgroundMethodInfluentialDavis, N., Xie, B., & Gurari, D. (2020) · Proceedings of the Association for Information Science and Technology · doi:10.1002/pra2.251 · Cited by 4
“With the intention of making this data openly available, this dataset was anonymized and filtered to remove any private or potentially private material (Gurari et al., 2018).”
BackgroundMethodInfluentialSelf-citationCOCO is “ALL” You Need for Visual Instruction Fine-tuning
Han, X., Wang, Y., Zhai, B., You, Q., & Yang, H. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687511 · Cited by 6
“Traditional vision-language models evaluation benchmarks include captions and VQAs, e.g. NoCaps [11], COCO [12], GQA [13], VQAv2 [14], VizWiz [15], etc.”
MethodMultimodal encoders and decoders with gate attention for visual question answering
Li, H., & Han, D. (2021) · Computer Science and Information Systems · doi:10.2298/csis201120032l · Cited by 4
“The VQA has the practically applied value in helping the blind [18][19] and image retrieval [16][27], etc.”
BackgroundArtificial Intelligence in Education
BAYRAKLI, V. K., & Demircioğlu, A. (2025) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-2612-2.ch002 · Cited by 4
Modality-specific Distillation
Jin, W., Sanjabi, M., Nie, S., Tan, L., Ren, X., & Firooz, H. (2021) · doi:10.18653/v1/2021.maiworkshop-1.7 · Cited by 4
“For example, visual question answering (VQA) requires answering questions about an image [2, 14, 16, 29] and models should incorporate information from the text and image sources to answer the questions.”
BackgroundConversations for Vision: Remote Sighted Assistants Helping People with Visual ImpairmentsPreprint
Lee, S., Reddie, M., Gurdasani, K., Wang, X., Beck, J., Rosson, M. B., & Carroll, J. M. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1812.00148 · Cited by 4
“There are drawbacks of VizWiz, though, with many of PVI’s questions being deemed unanswerable by remote workers [10].”
BackgroundGarg, K., Tang, X., Heo, J., Morgan, D. R., Gergle, D., Sudderth, E. B., & Piper, A. M. (2026) · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791309 · Cited by 3
“Although academic scholars and industry corporations have emphasized the pressing need for more disability-centered datasets [20, 38, 55, 79, 97, 111, 124], annotating these datasets with meaningful “ground truth” labels so that they can be used in benchmark studies and model evaluations such as…”
InfluentialAdversarial Training with OCR modality Perturbation for Scene-Text Visual Question Answering
Shen, Z., Luo, H., Li, S., & Li, T. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10688200 · Cited by 3
Seeing Eye Stretch: Robot-assisted Indoor Navigation for Blind or Low-Vision Individuals
Nevasekar, V., & Cabrera, M. E. (2025) · International Conference on Ubiquitous Robots (UR) · doi:10.1109/ur65550.2025.11078138 · Cited by 3
“We fine-tuned on the dataset [23] with 5000 examples for training, 1000 for validation, and 1000 for testing.”
Ground-V: Teaching VLMs to Ground Complex Instructions in Pixels
Zong, Y., Zhang, Q., An, D., Li, Z., Xu, X., Xu, L., Tu, Z., Xing, Y., & Dabeer, O. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02294 · Cited by 4
“Remarkably, we observe consistent performance improvements across all benchmarks, demonstrating that even generalist VLMs can benefit from Ground-V.”
Automatic Complexity Analysis of UML Class Diagrams Using Visual Question Answering (VQA) Techniques
Shehzadi, N., Ferzund, J., Fatima, R., & Riaz, A. (2025) · Software · doi:10.3390/software4040022 · Cited by 3
“Introducing the “VizWiz Grand Challenge,” Gurari et al. [10] developed Visual Question Answering (VQA) systems, especially for blind users.”
Da Yu: Toward ASV-Based Image Captioning for Waterway Surveillance and Scene Understanding
Guan, R., Ouyang, N., Xu, T. T., Liang, S., Dai, W., Sun, Y., Gao, S. S., Lai, S., Yao, S., Hu, X., et al. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3651269 · Cited by 5
Enhancing Few-Shot Vision-Language Classification With Large Multimodal Model Features
Mitra, C., Huang, B., Chai, T., Lin, Z., Arbelle, A., Feris, R., Karlinsky, L., Darrell, T., Ramanan, D., & Herzig, R. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00265 · Cited by 16
“(3) Vizwiz [19] is VQA dataset designed to progress research in vision systems to assist blind and vision-impaired individuals.”
Delta-LLaVA: Base-then-Specialize Alignment for Token-Efficient Vision-Language Models
Zamini, M., & Shukla, D. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00356 · Cited by 3
“Evaluation is performed on a diverse suite of benchmarks us-ing LMMs-Eval tool [3]: general VQA tasks ( V QA v 2 [9], GQA [12], VizWiz [10]); hallucination detection (POPE [18]); comprehensive reasoning benchmarks (MMBench [23], MME [20]); and image captioning (Nocaps [1] and Flicker30K[26]).”
Explaining CLIP's Performance Disparities on Data from Blind/Low Vision Users
Massiceti, D., Longden, C., Słowik, A., Wills, S., Grayson, M., & Morrison, C. P. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.01157 · Cited by 15
“VizWiz-Classification [8] contains 8,900 images from the original VizWiz dataset [25], a dataset of images taken by over 11,000 BLV users via a visual assistance mobile app [12].”
MethodGENIXER: Empowering Multimodal Large Language Model as a Powerful Data Generator
Zhao, H. H., Zhou, P., & Shou, M. Z. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73337-6_8 · Cited by 3
Fully Authentic Visual Question Answering Dataset from Online Communities
Chen, C., Liu, M., Codella, N., Li, Y., Yuan, L., & Gurari, D. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73195-2_15 · Cited by 11
Self-citationCan Commonsense Knowledge Improve CLIP's Performance in Cross-Domain VQA?
C, M. H. N., Jha, A., Abdar, M., & Banerjee, B. (2024) · Indian Conference on Computer Vision, Graphics and Image Processing · doi:10.1145/3702250.3702265 · Cited by 3
“Over the years, influential datasets like VQAv2 [12], VizWiz [14], Visual Genome [26], and Visual 7W [73] have emerged, driving research in this area.”
STEPS: Sequential Probability Tensor Estimation for Text-to-Image Hard Prompt Search
Qiu, Y., Wang, A., Li, C., Huang, H., Zhou, G., & Zhao, Q. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02667 · Cited by 3
“To ensure comprehensive evaluation, we conduct experiments on six diverse visual datasets: MS COCO [17], Flick30k [43], LAION [34], Celeb-A [18], textVQA [37], and VizWiz [13].”
Khan, A. S., Abbass, M. J., & Khan, A. (2025) · Sensors · doi:10.3390/s25195992 · Cited by 5
Taioli, F., Zorzi, E., Franchi, G., Castellini, A., Farinelli, A., Cristani, M., & Wang, Y. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01745 · Cited by 10
An Entropy Clustering Approach for Assessing Visual Question Difficulty
Terao, K., Tamaki, T., Raytchev, B., Kaneda, K., & Satoh, S. (2020) · IEEE Access · doi:10.1109/access.2020.3022063 · Cited by 3
“of attention in recent years because it has a large potential to impact many applications such as smart support for the visually impaired [3], providing instructions to autonomous robots [4], and for intelligent interaction between humans and machines [5].”
BackgroundUnderstanding Video Scenes through Text: Insights from Text-based Video Question Answering
Jahagirdar, S., Mathew, M., Karatzas, D., & Jawahar, C. V. (2023) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw60793.2023.00500 · Cited by 3
Toward Unsupervised Realistic Visual Question Answering
Zhang, Y., Ho, C.-H., & Vasconcelos, N. M. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01431 · Cited by 4
“Finally, the VQA datasets that support RVQA, such as VizWiz [17], are designed for a specific application domain, frequently containing images with few objects.”
BackgroundMethodInfluentialp-Laplacian Adaptation for Generative Pre-trained Vision-Language Models
Wu, H., Zhang, X., Xu, P. H., Liao, P., Yao, X., & Yu, B. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i6.28415 · Cited by 4
Oscar: Omni-scale robust contrastive learning for Text-VQA
Yue, J., Bi, X., & Chen, Z. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124785 · Cited by 3
Improving Accuracy and Generalizability via Multi-Modal Large Language Models Collaboration
Zhang, S., Mu, H., & Liu, T. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10651475 · Cited by 3
“For the general VQA task, follow the of setup [28], we utilize five benchmarks, including VQAv2 [35], OKVQA [36], GQA [37], ScienceQA (Image Set, zero-shot) [38], and VizWiz VQA (zero-shot) [39].”
MethodImageExplorer Deployment: Understanding Text-Based and Touch-Based Image Exploration in the Wild
Xu, A., Cai, M., Hou, D., Chang, R.-C., & Guo, A. (2024) · International Web for All Conference (W4A) · doi:10.1145/3677846.3677861 · Cited by 3
Self-citationMedina, M. J., Thoo, Y.-J., Baudet, C., Froehlich, J. E., Ruffieux, N., & Lalanne, D. (2025) · ACM Transactions on Accessible Computing · doi:10.1145/3726531 · Cited by 5
“Six papers ( N = 6) provided datasets (e.g., [7, 42]), with one notable example being Gurari et al. [67], who collected over 30,000 visual questions from blind individuals using mobile phones, each paired with spoken questions and crowdsourced answers.”
@BENCH: Benchmarking Vision-Language Models for Human-centered Assistive Technology
Jiang, X., Zheng, J., Liu, R.-P., Li, J., Zhang, J., Matthiesen, S., & Stiefelhagen, R. (2024) · IEEE Workshop/Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61041.2025.00387 · Cited by 7
“By posing questions, they can experience and comprehend the un-seen world around them, providing them with a novel and enriching experience.”
InfluentialDiViCo: Disentangled Visual Token Compression for Efficient Large Vision-Language Model
Wang, X. Y., Pan, Z., Chen, H., & Zhu, W. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3567138 · Cited by 3
“[57]: The Vizwiz benchmark consists of over 31 , 000 visual questions originating from a group of visually impaired people each of whom takes a picture using a mobile phone and records a spoken question about it, together with 10 crowdsourced answers per visual question.”
Le, T., Huy, N. T., & Minh, N. L. (2020) · International Conference on Knowledge and Systems Engineering/Proceedings · doi:10.1109/kse50997.2020.9287539 · Cited by 4
Ground-truth or DAER: Selective Re-query of Secondary Information
Lemmer, S. J., & Corso, J. J. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00074 · Cited by 3
“For these tasks, correct and incorrect seeds can be classified in a manner similar to curated VQA datasets [37, 19] while additional error can be regressed directly.”
BackgroundLook and Answer the Question: On the Role of Vision in Embodied Question Answering
Ilinykh, N., Emampoor, Y., & Dobnik, S. (2022) · International Conference on Natural Language Generation · doi:10.18653/v1/2022.inlg-main.19 · Cited by 3
“These problems combined make the task harder and bridge it with the likes of captioning of images taken by visually impaired people (Gurari et al., 2018) instead of VQA where images are fixed and taken in perfect conditions to answer the question.”
BackgroundEaSe: A Diagnostic Tool for VQA based on Answer Diversity
Jolly, S., Pezzelle, S., & Nabi, M. (2021) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2021.naacl-main.192 · Cited by 3
“0 (Goyal et al., 2017) and VizWiz (VW; Gurari et al., 2018).”
BackgroundTran, K. V., Phan, H. P., Nguyen, K. V., & Nguyen, N. L. T. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4611463 · Cited by 6
Cripping AI: Reimagining AI Through Lived Disability Experiences
Tang, X., Lin, T.-A., Li, J., & Wu, S. (2026) · Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3805689.3806744 · Cited by 2
“…neurodivergence is conceptualized such as in datasets [163] and evaluation measurements [67], engage with neurodivergent communities to understand their perspectives [62, 230], and transform existing practices to better support neurodivergent expression, from design goals to evaluation [236, 268].”
InfluentialHuman vs LLM-Generated Alt Text for STEM Image Accessibility: A Comparative Study
Cardia, M., Angileri, L., Poggianti, C., & Leporini, B. (2026) · Proceedings of the 23rd International Web for All Conference · doi:10.1145/3800424.3800457 · Cited by 2
“Over the past decade, progress in image captioning has been driven by the availability of large-scale datasets such as Flickr8k/30k [11], MS COCO [17], and VizWiz [9].”
Metadata
- Authors
- Gurari, D., Li, Q., Stangl, A. J., Guo, A., Lin, C., Grauman, K., Luo, J., & Bigham, J. P.
- Affiliations
- The University of Texas at Austin (US)
- University of Science and Technology of China (CN)
- University of Colorado Boulder (US)
- Carnegie Mellon University (US)
- University of Rochester (US)
- Venue
- 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
- Publisher
- IEEE
- Published
- 2018-06
- Pages
- 3608-3617
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W2788643321
- Cited by
- 621
- References
- 45
- Citations by year
- 2026: 92 · 2025: 133 · 2024: 108 · 2023: 65 · 2022: 56 · 2021: 66 · 2020: 55 · 2019: 29 · 2018: 17
- FWCI
- 12.58
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Science Foundation (1704309, 1514118, 1755593)
- Microsoft Research
- Baidu
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image Retrieval and Classification Techniques
- Keywords
- visual question answering, blind people, assistive technology, low-quality images, answerability
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
