SparseVSR：轻量级且抗噪声的视觉语音识别
SparseVSR: Lightweight and Noise Robust Visual Speech Recognition
AI总结：
本文提出SparseVSR，通过基于幅度的剪枝技术构建轻量级视觉语音识别模型，在LRS3数据集上以10%稀疏度取得最优结果，并证明稀疏网络比稠密网络更抗视觉噪声。
摘要：
深度神经网络的最新进展在视觉语音识别中取得了前所未有的成功。然而，当前方法与它们在资源受限设备上的部署之间仍存在显著差距。在这项工作中，我们探索了不同的基于幅度的剪枝技术，以生成一种轻量级模型，该模型比其稠密模型等价物取得更高性能，尤其是在存在视觉噪声的情况下。我们的稀疏模型在LRS3数据集上以10%稀疏度取得了最先进的结果，并在高达70%稀疏度时优于稠密等价物。我们在7种不同的视觉噪声类型上评估了50%稀疏度的模型，与稠密等价物相比，整体WER绝对改善超过2%。我们的结果证实，稀疏网络比稠密网络对噪声更具抵抗力。
英文摘要：
Recent advances in deep neural networks have achieved unprecedented success in visual speech recognition. However, there remains substantial disparity between current methods and their deployment in resource-constrained devices. In this work, we explore different magnitude-based pruning techniques to generate a lightweight model that achieves higher performance than its dense model equivalent, especially under the presence of visual noise. Our sparse models achieve state-of-the-art results at 10% sparsity on the LRS3 dataset and outperform the dense equivalent up to 70% sparsity. We evaluate our 50% sparse model on 7 different visual noise types and achieve an overall absolute improvement of more than 2% WER compared to the dense equivalent. Our results confirm that sparse networks are more resistant to noise than dense networks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
