{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/dataset/gotcha","name":"GOTCHA","full_name":null,"description_markdown":"We release the dataset for non-commercial research. Submit requests <a href=\"https://forms.gle/6WPEGNWbYoEe6bte8\" target=\"_blank\">here</a>.\r\n\r\nThe dataset consists of three parts: Original Samples, Deepfakes, and the Fidelity Score Dataset.\r\n\r\nThe dataset is accessible via Box.com, which offers high availability worldwide and allows for access tracking.\r\n\r\n### Dataset Statistics:\r\n\r\n**Number of folder (or videos):**\r\n- Original : 675 (673k frames with 3 frame sizes each)\r\n- Fake (DFL) : 26,085 (5.0M frames)\r\n- Fake (Adaptive - DFL) : 7,215 (1.3M frames)\r\n- Fake (FSGAN) : 25,433 (4.6M frames)\r\n- Fake (LIA) : 25,662 (5.0M frames)\r\n\r\n**DFL quantized ready-to-use models** : 60\r\n\r\n**Fidelity Score Dataset:**\r\n- Fake : 49,603 videos (32 frames each, 1.35M frames total)\r\n- Real : 816 videos (32 frames each, 22k frames total)\r\n\r\n**Human evaluation dataset:** 200 videos\r\n\r\n\r\n### Original Samples\r\n\r\nWhile the paper focuses on 8 challenges, we collected 13 in total, which we share in full. The data is provided as frames extracted from raw HD videos recorded at 60 fps. We downsample the frames to approximately 10 fps for ease of use without any expected loss in generation fidelity. The images inside each zip file are face-only. We provide three convenient sizes: 224 x 224, 512 x 512, and 1024 x 1024 pixels.\r\n\r\nEach zip file contains folders numbered according to the following challenges/tasks:\r\n\r\n0. The camera moves around for about two minutes, capturing the face from different angles while the participant sits still.\r\n1. The participant rotates their head from side to side, then looks up and down. Each side should be held for about 5 seconds, rotating as comfortably as possible, totaling around 25 seconds.\r\n2. The participant covers their eyes with a hand, followed by covering the left half, the right half, and finally, the lower half of the face.\r\n3. The participant puts on the provided sunglasses and then takes them off.\r\n4. The participant wears the provided clear glasses, ensuring they reflect a lamp light shining. After setting up the reflection, recording starts. Finally, the participant removes the glasses.\r\n5. The participant moves a green cloth in front of their face\r\n6. The participant puts on a face mask and counts from 1 to 10 out loud. Then, they remove the facemask.\r\n7. The participant presses a finger against a cheek.\r\n8. The participant sticks out a (small) portion of their tongue.\r\n9. The participant laughs for 10 seconds, then frowns, as if angry, for another 10 seconds.\r\n10. The participant slowly stands up and then sits back down.\r\n11. The participant counts from 1 to 50 with distinct breaks between each utterance.\r\n12. The room light is dimmed, and a flashing light is kept on the face for 10 seconds.\r\n\r\n**TODO: Verify each challenge index above corresponds to that challenge in the zips.**\r\n\r\nNote: A few challenges are missing from a handful of participants. \r\n\r\n### Fakes\r\n\r\nUsing 47 participants, we create deepfakes between all 2,209 (47 x 47) combinations (including self deepfake), using three deepfake generators — DeepFaceLab, FSGANv2, and LIA. \r\n\r\n1. **DFL (DeepFaceLab)**: Notorious for generating hyper-realistic deepfakes, this pipeline serves as a baseline for in-the-wild deepfake videos. For our study, we trained *individual DFL deepfake generators for each participant* using their `no challenge' videos. These videos records the participants in a range of frontal angles while they sit naturally, aiming to mimic the kind of data readily accessible online for non-celebrity individuals. Training continued until convergence for approximately around 300,000 iterations. \r\n    \r\n    **DFL (Adaptive Adversary)**: We utilized heavily trained celebrity deepfake generators provided by the community as an adaptive adversary. Such models are trained on upwards of 2M iterations.\r\n    \r\n    The DFL models trained by us and the ones provided by the community  are available under *dfm_models* folder, including the hyperparameters in states.dat file. The models can be loaded directly into Deepfacelive to generate deepfakes.\r\n    \r\n2. **FSGAN (Face Swapping Generative Adversarial Network)**: This corresponds to the second version of FSGAN. Similar to LIA, this model is also target-agnostic, hence we utilized a pre-trained model made available by the authors for our study. Access its release at https://github.com/YuvalNirkin/fsgan\r\n3. **LIA (Latent Image Animator)**: This pipeline is a facial reenactment method outlined in. Given its target-agnostic nature—meaning it does not require specific target data during inference—we employed a pre-trained model for our experiments. Access its release at https://github.com/wyhsirius/LIA\r\n\r\nAs a rule of thumb, An imposter outer face and target is the inner face, in case of faceswaps. \r\n\r\nWhile Box offers quick zipping of multiple files, we have placed the files into zips (<10 GB each), for convenient access. Each zip under fakes (except Adaptive Adversary) is organized in path : “target/challenge/imposter/*.jpg”. For Adaptive Adversary, the order is opposite, i.e., “imposter/challenge/target/*.jpg”\r\n\r\n### Fidelity Score Dataset (or a Snapshot of the full dataset)\r\n\r\nThe training subset contains the following subselections:\r\n\r\n- Original and deepfake videos with four challenges, one from each category and ‘no challenge’.\r\n- 32 frames per sample,\r\n- 35 out of 47 target identities, and\r\n- Deepfake videos created using only DFL.\r\n\r\nThe validation dataset consists of the remaining 12 out of 47 target identities, while test one includes the celebrity deepfakes. The dataset is the smallest snapshot of the full dataset. The code for loading the dataset and training a fidelity score function can be accessed at [https://github.com/mittalgovind/gotcha-deepfakes](https://github.com/mittalgovind/gotcha-deepfakes).\r\n\r\n### Human evaluation\r\n\r\nWe release the instruments we used for human evaluation at [https://app.gorilla.sc/openmaterials/693684](https://app.gorilla.sc/openmaterials/693684) (with active preview).","description_withheld":null,"homepage":"https://github.com/mittalgovind/GOTCHA-Deepfakes","introduced_date":"2022-10-12","introduced_date_note":null,"introduced_by":{"paper":"/paper/gotcha-a-challenge-response-system-for-real","title":"GOTCHA: Real-Time Video Deepfake Detection via Challenge-Response","first_author":"Govind Mittal","url":null},"license":null,"modalities":[{"name":"Images","url":"/datasets/modality/images"},{"name":"Videos","url":"/datasets/modality/videos"},{"name":"Speech","url":"/datasets/modality/speech"}],"tasks":[{"name":"DeepFake Detection","url":"/task/deepfake-detection","datasets_with_task":"/datasets/task/deepfake-detection"},{"name":"Human Detection of Deepfakes","url":"/task/human-detection-of-deepfakes","datasets_with_task":"/datasets/task/human-detection-of-deepfakes"}],"languages":[{"name":"English","url":"/datasets/language/english"}],"variants":["GOTCHA"],"data_loaders":[],"num_papers_in_archive":2,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[],"papers":{"order":"date (newest first), then slug","population":"every paper with a leaderboard row on this dataset's benchmarks (the benchmark-backed subset): the archive's own papers-using-this-dataset list was never published, so this is not that list; num_papers_in_archive is the archive's own count","page":1,"pages_in_order":1,"rows_per_page":100,"shown":0,"of":0,"counts":{"papers_with_a_benchmark_row":0,"with_a_code_link":0,"where_syntology_ran_a_sample":0,"not_listed_spam_title":0,"listed":0,"listed_where_code_ran":0,"where_syntology_ran_a_sample_split":{"with_a_run_with_no_instrument_failure":0,"every_run_a_failure_of_syntologys_instrument":0,"listed_with_a_run_with_no_instrument_failure":0,"listed_every_run_a_failure_of_syntologys_instrument":0,"filter":{"states":["a run with no instrument failure","any run, instrument failures included"],"default":"a run with no instrument failure","note":"on the 'only where code ran' pages the default hides, in the browser, the rows where every run was a failure of Syntology's instrument; the second state shows them again. Rows are hidden, never re-ordered; these twins list every row"}},"definition":"distinct papers with at least one leaderboard row on this dataset's benchmarks; 'where Syntology ran a sample' counts papers with at least one harvested sample that ran, which is not a correctness claim"},"items":[],"full_list":"papers_with_a_benchmark_row: every listed paper, in this order"},"papers_with_a_benchmark_row":[],"papers_where_code_ran":0,"papers_where_code_ran_first_page":null,"papers_where_code_ran_caption":"We ran code from the paper's repository; we did not run it on this dataset or check it against this dataset's benchmarks.","pages":[],"row_route":"each paper's syntology object carries sample_list (this site's paper page, #ran), syntology_url (syntology.ai's page for the paper, only when that route probed live) and mcp (the agent call that lists the paper's harvested samples; not a verdict on the paper)","developers":"https://syntology.ai/developers","syntology_totals":{"read_at":"2026-10-05T00:36:11+00:00","papers_with_samples":0,"samples_harvested":0,"samples_ran":0,"samples_constructed":0,"samples_ran_checked":0,"samples_ran_instrument_failed":0,"samples_unverified":0,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"papers_with_a_sample_that_ran":0,"papers_with_a_run_with_no_instrument_failure":0,"papers_where_every_run_was_an_instrument_failure":0,"note":"the per-paper counts of every listed paper (all pages), summed; not a rate"},"key_notes":{"samples_ran_checked":"legacy name, kept unchanged so existing readers do not break: it counts the samples that ran with no instrument failure (honoured, violated, and ran with no contract checked); it does not mean a contract was checked, and the pages print it as 'K with no instrument failure', not 'K checked'","samples_constructed":"a sub-count of the samples that ran, never subtracted from them and never a failure: an executed sample whose run returned an instance of its own class (fixture_out_type equals the entry name): the run built an object and did not compute a result (Syntology's RAN record, counts.constructed)"},"since_archive_judged":{"papers":0,"papers_placed":0,"judged_by":{"model":"a model","models_as_recorded":[],"from":"lb_rows_run_verified.jsonl: each record's `model` field (the extractor's call that wrote decisions[])"},"note":"a model's judgement from the paper's own tables that they report on this leaderboard, not a result: no value from a paper that was not placed is printed or carried here; papers counts each paper once over these leaderboards","backfill":"8,886 of 18,637 newer papers checked so far (733 more had nothing to place; 41 could not be read)","backfill_complete":false,"leaderboards":[]},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks.","record_sha256":"633437b730cbec7637850f144c4d0feb4eb0d04c32a906a5858615d36da47cf5","record_changed_at":"2026-10-05","record_changed_at_basis":"changed"}