Multi-Region Probabilistic Histograms for Robust and Scalable Identity Inference

By Conrad Sanderson, Brian C. Lovell. International Conference on Biometrics, 2009

334 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/d3306e3e9fb1b0af59f9bf1de0952b995e1440cd

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

In Computer Vision and Pattern Recognition

Citations in the corpus

Object Recognition with and without Objects (2016, methodology): “For face matching, [Sanderson and Lovell, 2009] developed methods only on the cropped out faces to alleviate the possible correlations between faces and their backgrounds.”
Patch-based probabilistic image quality assessment for face selection and improved video-based face recognition (2011, methodology): “Each face set was represented by an average MRH signature; LBP feature extraction was not used as it isn’t suitable for feature averaging.”
DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection (2020, methodology): “This database comprises 620 fake videos of 32 subjects from the VidTIMIT database [48].”
Armadillo: An Open Source C++ Linear Algebra Library for Fast Prototyping and Computationally Intensive Experiments (2010, methodology): “The library has been successfully used for the development of several algorithms, such as [12, 13, 15].”
Deep Audio-visual Learning: A Survey (2020, background): “Some of the datasets only contain videos of speakers that utter the given sentences; these datasets include GRID [119] , TCD TIMIT [121] , and VidTIMIT [122] .”
FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Dataset (2021, background): “UADF consists of only a few (98) videos; meanwhile, Deepfake TIMIT contains audio along with the video as well, except the audios are real and not synthesized or fake.”
Deepfakes generation and detection: state-of-the-art, open challenges, countermeasures, and way forward (2021, methodology): “drops, however, for the VidTIMIT [115] and AMI [116] datasets due to fewer training samples for each person in both of these datasets over the GRID dataset.”
Exposing DeepFake Videos By Detecting Face Warping Artifacts (2018, methodology): “video set has 32 subjects, where each subject has 10 videos with faces swapped. Each video is 512 384 and lasts ˘ 4 seconds. The original videos of corresponding 32 subjects are from VidTIMIT dataset [29]. We select subset of each subject from original dataset VidTIMIT and all fake videos from Deepf”
A Continual Deepfake Detection Benchmark: Dataset, Methods, and Essentials (2022, methodology): “Deepfake-TIMIT [40] VidTIMIT dataset [66] Known Deepfake tech ✗”
Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics (2019, methodology): “DF-TIMIT : The DeepFake-TIMIT dataset [25] includes 640 DeepFake videos generated with faceswap-GAN [3] and based on the Vid-TIMIT dataset [43].”
Graph embedding discriminant analysis on Grassmannian manifolds for improved image set matching (2011, background): “This is mainly driven by the need for superior discrimination accuracy as well as increased robustness to practical issues such as pose variations, misalignment and varying environmental conditions (for example, as present in realistic face recognition scenarios [21]).”
Labeled Faces in the Wild: A Survey (2016, background): “- Eigenfaces, 1991 [99], Nowak, 2007 [74], Multi-Region histograms, 2009 [81], Learning-based descriptor, 2010 [26]”
Conditional generative adversarial nets for convolutional face generation (2015, methodology): “Figure 3: Random samples from the Labeled Faces in the Wild dataset [14] ((a), (b), and (c)) and their equivalents (d), (e), and (f) in the LFWcrop [20] dataset.”
Multiple instance classification: Review, taxonomy and comparative study (2013, methodology): “Finally, Speaker was created by us based on the audio database available in the website [43] and introduced in the paper [44].”
Two-branch Recurrent Network for Isolating Deepfakes in Videos (2020, methodology): “, DF-TIMIT [33,34] using 32 subjects selected from the VidTIMIT [53] database with the intent of studying the weaknesses of face detection and recognition technology, or UADFV [36] that offers around 50 bona fide and 50 fake videos.”