HyperAI超神经

首页算力平台文档资讯论文教程数据集百科 SOTA LLM 模型天梯 GPU 天梯顶会

中文

HyperAI超神经

Lipreading On Lrw 1000

评估指标

Top-1 Accuracy

评测结果

各个模型在此基准测试上的表现结果

		Paper Title	Repository
SyncVSR (Word Boundary)	58.2	SyncVSR: Data-Efficient Visual Speech Recognition with End-to-End Crossmodal Audio Token Synchronization
3D-ResNet + Bi-GRU + MixUp + Label Smooth + Cosine LR (Word Boundary)	55.7%	Learn an Effective Lip Reading Model without Pains
3D Conv + ResNet-18 + MS-TCN + Multi-Head Visual-Audio Memory	53.8	Distinguishing Homophenes Using Multi-Head Visual-Audio Memory for Lip Reading
3D Conv + ResNet-18 + Bi-GRU + Visual-Audio Memory	50.82%	Multi-modality Associative Bridging through Memory: Speech Sound Recollected from Face Video	-
3D-ResNet + Bi-GRU + MixUp + Label Smooth + Cosine LR	48.3%	Learn an Effective Lip Reading Model without Pains
3D Conv + ResNet-18 + Bi-GRU (Face Cutout)	45.24%	Can We Read Speech Beyond the Lips? Rethinking RoI Selection for Deep Visual Speech Recognition
DFTN	41.93%	Deformation Flow Based Two-Stream Network for Lip Reading
GLMIM	38.79%	Mutual Information Maximization for Effective Lip Reading
PCPG	38.7%	Pseudo-Convolutional Policy Gradient for Sequence-to-Sequence Lip-Reading	-

0 of 9 row(s) selected.

Lipreading On Lrw 1000 | SOTA | HyperAI超神经