An audio-visual corpus for speech perception and automatic speech recognition.

By Martin P. Cooke, Jon Barker, S. Cunningham, Xu Shao. 2006

1,232 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Experimental and Cognitive Psychology, 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 Cognitive Neuroscience

In Experimental and Cognitive Psychology

In Signal Processing

Citations in the corpus

Phased LSTM: Accelerating Recurrent Network Training for Long or Event-based Sequences (2016, methodology): “For this task, we use the GRID dataset [8].”
Audio-Visual Scene Analysis with Self-Supervised Multisensory Features (2018, methodology): “Finally, we also transferred (without retraining) our learned models to the GRID dataset [73], a labrecorded dataset in which people speak simple phrases in front of a plain background, finding a similar relative ordering of the methods.”
The DIRHA-ENGLISH corpus and related tasks for distant-speech recognition in domestic environments (2015, background): “…on Acoustic Echo Cancellation (AEC), to suppress known interferences dif-fused in the environment (e.g., played music); • The DIRHA-GRID corpus [25], a multi-microphone multi-room simulated data set that derives from contaminating the GRID corpus [26] of short commands in the English language.”
End-to-End Audiovisual Speech Recognition (2018, background): “The number of words, 500, is also much higher than existing lipreading databases used for word recognition, which typically contain 10 to 50 words [20, 21, 17].”
Lip Movements Generation at a Glance (2018, methodology): “We verify the delayed correspondence problem between audio and visual information by designing a case study on 3260 videos randomly sampled from the GRID dataset.”
Hierarchical Cross-Modal Talking Face Generation With Dynamic Pixel-Wise Loss (2019, methodology): “We evaluate our model along with state-of-the-art methods on several popular datasets (e.g., GRID [6], LRW [5], VoxCeleb [24] and TCD [13]).”
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] .”
ReVISE: Self-Supervised Speech Resynthesis with Visual Input for Universal and Generalized Speech Enhancement (2022, background): “The model demonstrate strong re-sults on two constrained datasets (LRW [8] and GRID [11]) which contain isolated words and ﬁxed-length sentences with limited word choices, respectively.”
Lip Reading Sentences in the Wild (2016, methodology): “This reports strong speaker-independent performance on the constrained grammar and 51 word vocabulary of the GRID dataset [11].”
What comprises a good talking-head video generation?: A Survey and Benchmark (2020, background): “In the GRID dataset (Cooke et al. 2006), there are 33 speakers, front-facing the camera and each uttering 1000 short phrases, containing six words randomly chosen from a limited dictionary (51 words).”
A Probabilistic Interaction Model for Multipitch Tracking With Factorial Hidden Markov Models (2011, methodology): “2) Two male and two female speakers (abbreviated as MA1, MA2 and FE1, FE2, respectively), were selected from the GRID database [24], and 500 English sentences were selected per speaker.”
Large-Scale Visual Speech Recognition (2018, methodology): “y predicting character sequences. The model combined spatiotemporal convolutions with gated recurrent units (GRUs) and was trained using the CTC loss function. LipNet was evaluated on the GRID corpus [41], a limited grammar and vocabulary dataset consisting of 28 hours of 5-word sentences, where it ”
Lip Reading in Profile (2017, neutral): “Name Type View Vocab # Utterances OuluVS2 [2] Fixed phrases 0◦ - 90◦ 10 3,640 GRID [11] Phrases 0◦ 51 33,000 LRW [9] Words 0◦ - 30◦ † 500 500,000 LRS [8] Sentences 0◦ - 30◦ † 17,428 118,116 MV-LRS Sentences 0◦ - 90◦ † 14,960 74,564 Table 1: Comparison of existing datasets.”
Single-channel mixed speech recognition using deep neural networks (2014, methodology): “The speech data of the challenge task is drawn from GRID corpus [17].”
The Sweet-Home speech and multimodal corpus for home automation interaction (2014, background): “For instance, in the context of meeting room analysis (Mccowan, 2003) or of home automation order recognition from distant speech (Cooke et al., 2006; Christensen et al., 2010).”
A Lip Sync Expert Is All You Need for Speech to Lip Generation In the Wild (2020, background): “Several works [5, 26, 28] train on datasets with a limited set of words such as GRID [10] (56 words), TIMIT [14] and LRW [8] (1000 words) which significantly hampers a model from learning the vast diversity of phoneme-viseme mappings in real videos [18].”