Lip Reading in Profile

By Joon Son Chung, Andrew Zisserman. 2017

146 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

VoxCeleb2: Deep Speaker Recognition (2018, methodology): “This is done by using a multi-view adaptation [30] of ‘SyncNet’ [31, 32], a twostream CNN which determines the active speaker by estimating the correlation between the audio track and the mouth motion of the video.”
Spot the conversation: speaker diarisation in the wild (2020, methodology): “The audio-visual method requires three key thresholds – cosine distance for face clustering, SyncNet confidence for active speaker detection, and cosine distance for speaker identification.”
An Overview of Deep-Learning-Based Audio-Visual Speech Enhancement and Separation (2020, methodology): “[41], generic YouTube videos [37], [189], TED talks [9], [55],”
Deep Audio-Visual Speech Recognition (2018, methodology): “The networks are first trained on the frozen features of the pre-train sets from MV-LRS, LRS2-BBC and LRS3-TED.”
The Conversation: Deep Audio-Visual Speech Enhancement (2018, methodology): “The model is trained on two datasets: the first is the BBCOxford Lip Reading Sentences 2 (LRS2) dataset [34, 39], which contains thousands of sentences from BBC programs such as Doctors and EastEnders; the second is VoxCeleb2 [40], which contains over a million utterances spoken by over 6,000 differ”
Large-Scale Visual Speech Recognition (2018, methodology): “Finally, our dataset, LSVSR, is an order of magnitude larger than existing sentence-level visual speech recognition datasets ([2, 3, 28]; Table 1), and the content is more diverse and varied as it consists of general YouTube videos rather than solely news broadcasts, conference talks, or recording s”
Deep Audio-visual Learning: A Survey (2020, unknown): “Classes Length* Year GRID [119] Lab 34 33,000 2006 Lombard Grid [120] Lab 54 54,000 2018 TCD TIMIT [121] Lab 62 2015 Vid TIMIT [122] Lab 43 2009 RAVDESS [123] Lab 24 2018 SEWA [124] Lab 180 2017 Speech OuluVS [125] Lab 20 1000 2009 OuluVS2 [126] Lab 52 3640 2016 Voxceleb [127] Wild 1,251 154,516 201”
Talking Face Generation by Adversarially Disentangled Audio-Visual Representation (2018, methodology): “In [8], they further fixed the sync image feature as the pretraining for lip reading, but the two tasks are still separate from each other.”