{"id":13690024,"url":"https://github.com/kushalkafle/DVQA_dataset","last_synced_at":"2025-05-02T06:31:36.065Z","repository":{"id":99269881,"uuid":"136659718","full_name":"kushalkafle/DVQA_dataset","owner":"kushalkafle","description":"DVQA Dataset: A Bar chart question answering dataset presented at CVPR 2018","archived":false,"fork":false,"pushed_at":"2019-06-24T01:16:18.000Z","size":13,"stargazers_count":32,"open_issues_count":1,"forks_count":2,"subscribers_count":4,"default_branch":"master","last_synced_at":"2024-11-12T15:43:16.775Z","etag":null,"topics":["bar-chart","cvpr2018","dataset","deep-learning","question-answering","vqa"],"latest_commit_sha":null,"homepage":null,"language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/kushalkafle.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2018-06-08T19:43:35.000Z","updated_at":"2024-10-07T21:47:43.000Z","dependencies_parsed_at":"2023-07-17T23:01:01.442Z","dependency_job_id":null,"html_url":"https://github.com/kushalkafle/DVQA_dataset","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kushalkafle%2FDVQA_dataset","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kushalkafle%2FDVQA_dataset/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kushalkafle%2FDVQA_dataset/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kushalkafle%2FDVQA_dataset/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/kushalkafle","download_url":"https://codeload.github.com/kushalkafle/DVQA_dataset/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":251998527,"owners_count":21677993,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["bar-chart","cvpr2018","dataset","deep-learning","question-answering","vqa"],"created_at":"2024-08-02T16:00:42.105Z","updated_at":"2025-05-02T06:31:36.059Z","avatar_url":"https://github.com/kushalkafle.png","language":null,"funding_links":[],"categories":["Papers"],"sub_categories":["Natural Language for Visual Reasoning"],"readme":"# DVQA\nThis repository provides the images, metadata and question-answer pairs described in the paper:\n\n**[DVQA: Understanding Data Visualizations via Question Answering](https://arxiv.org/abs/1801.08163)**\n \u003cbr\u003e\n \u003ca href=\"https://kushalkafle.com/\" target=\"_blank\"\u003eKushal Kafle\u003c/a\u003e,\n \u003ca href=\"https://research.adobe.com/person/brian-price/\" target=\"_blank\"\u003eBrian Price\u003c/a\u003e,\n\u003ca href=\"https://research.adobe.com/person/scott-cohen/\" target=\"_blank\"\u003eScott Cohen\u003c/a\u003e,\n\u003ca href=\"http://www.chriskanan.com/\" target=\"_blank\"\u003eChristopher Kanan\u003c/a\u003e\n\n\nTo be presented at \u003ca href=\"http://cvpr2018.thecvf.com/\" target=\"_blank\"\u003eCVPR 2018\u003c/a\u003e\n\n\n\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"https://kushalkafle.com/images/dvqa.png\" width=\"450px\"\u003e\n\u003c/div\u003e\n\n\nPlease cite the following if you use the DVQA dataset in your work:\n```\n@inproceedings{kafle2018dvqa,\n  title={DVQA: Understanding Data Visualizations via Question Answering},\n  author={Kafle, Kushal and Cohen, Scott and Price, Brian and Kanan, Christopher},\n  booktitle={CVPR},\n  year={2018}\n}\n```\n\nA live demo of our `SANDY` algorithm as described in the paper above can be found in this \u003ca href='http://askimage.org'\u003eurl\u003c/a\u003e\n# Download Links\n\n#### Images\n\nDownload images using this \u003ca href='https://drive.google.com/file/d/1iKH2lTi1-QxtNUVRxTUWFvUvRHq6HAsZ/view?usp=sharing'\u003eurl\u003c/a\u003e. The images are all in the same folder and are named as\n```\nbar_{split}_xxxxxxxx.png\nwhere, \nxxxxxxxx = image_id padded (right justified) to length of 8 characters\nsplit = train, val_easy, or val_hard\n```\nThe images expand to about 6.5 GB.\n\n#### Question Answer Pairs\nThe question-answer pair can be downloaded from this  \u003ca href='https://drive.google.com/file/d/1VKYd3kaiCFziSsSv4SgQJ2T5m7jxuh5u/view?usp=sharing'\u003eurl\u003c/a\u003e. It consists of three files, one each for three different splits of the dataset named as `{split}_qa.json` It consists the following fields:\n\n```\nimage: The image filename which the given question-answer pair applies to\nquestion: Question\nanswer: Answer to the Questions. Remember that (cardinal numbers (1,2,3...) are used when \n\tthe number denotes the value and words (one,two,three...) are used to denote count\nquestion_type: Denotes whether the question is structure, data or reasoning type\nbbox_answer: If the answer is a text in the bar_chart, bounding box in form of [x,y,w,h], else []\nquestion_id: Unique question_id associated with the question\n```\nThe question-answer pairs expand to about 750 MB.\n\n#### Bar-chart metadata\nIn addition to question-answers, we also provide detailed annotations of every object in the bar-chart that can serve as either the source of additional supervision (à la our SANDY and MOM model) or use it to do additional analysis of your algorithm's performance. \n\nMetadata for the bar-charts can be downloaded using this \u003ca href='https://drive.google.com/file/d/1vBz8Ji4TMY7rzTL2_DJCTUEyWR7l16W6/view?usp=sharing'\u003eurl\u003c/a\u003e. It consists of three files, one each for three different splits of the dataset named as `{split}_metadata.json` It consists the following fields:\n\n```\nimage: The image filename which the given metadata applies to\nbars:\n\tbboxes: Bounding boxes for different bars (number_of_bars x number_of_legends x 4)\n    \tnames: Names for each bar in the form (number_of_bars x number_of_legends)\n\tcolors: Color of each bar (number_of_bars x number_of_legends)\n\ntexts:\n\ttext: The string of the text-block in the bar-chart\n    \ttext_function: The function of text (e.g., title, legend, etc)\n    \tbbox: The bounding box surrounding the text-block\n\ntable: Underlying table used to create the chart saved in the following format.\n\n\tsingle row charts:\n\t\tC_1 \tC_2 \tC_3\t...\tC_N\n\t\t-------------------------------------\n\t\tV_1\tV_2\tV_3\t... \tV_N\n\t\t\n\tmulti row charts:\n\t\t\n\t\tNone |\tC_1 \tC_2 \tC_3\t...\tC_N\n\t\t-----|---------------------------------------\n\t\tR_1  |\tV_11\tV_21\tV_31\t... \tV_N1\n\t\tR_2  |\tV_12\tV_22\tV_32\t... \tV_N2\n\t\t...  |\t...\t...\t... \t... \t...\n\t\tR_M  |\tV_1M\tV_2M\tV_3M\t... \tV_NM\n\t\n```\nSince numpy arrays are not supporte by JSON, the tables are saved as nested lists. Converting them to numpy array, e.g., `table = np.array(metadata['table'])` might provide easier access to the elements, e.g., for multi-row charts, `table[1:,1:]` contains the numeric data, `table[1:,0]` contains the row names and `table[0,1:]` contain the column names.\n\nThe annotations expand to about 800 MB.\n# Contact\nFeel free to contact us (contact details on the paper PDF) about any questions, suggestions or comments about either the dataset or the methods used in the paper.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkushalkafle%2FDVQA_dataset","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fkushalkafle%2FDVQA_dataset","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkushalkafle%2FDVQA_dataset/lists"}