增强情感信号：面向鲁棒语音情感识别的数据高效深度学习
Amplifying Emotional Signals: Data-Efficient Deep Learning for Robust Speech Emotion Recognition
AI总结：
针对语音情感识别中数据有限的难题，本文结合迁移学习与数据增强技术训练多种模型，其中ResNet34在RAVDESS与SAVEE组合数据集上取得新基准，验证了该方案应对数据稀缺的有效性。
摘要：
语音情感识别（Speech Emotion Recognition, SER）是人机交互领域一项重要且长期存在的挑战。尽管深度学习已推动口语处理取得进展，但在有限数据集上实现高性能仍是关键障碍。本文针对这一问题，开发并评估了一套用于人类语音自动情感分类的机器学习模型，包括支持向量机（Support Vector Machines, SVMs）、长短期记忆网络（Long Short-Term Memory networks, LSTMs）和卷积神经网络（Convolutional Neural Networks, CNNs）。我们证明，通过策略性地运用迁移学习和创新的数据增强技术，即使数据集规模相对较小，我们的模型仍能取得出色性能。我们性能最优的模型为ResNet34架构，在RAVDESS和SAVEE组合数据集上创下新的性能基准，准确率达66.7%，F1值达0.631。这些结果凸显了利用预训练模型和数据增强克服数据稀缺的显著优势，从而为构建更鲁棒、泛化性更强的SER系统铺平了道路。
英文摘要：
Speech Emotion Recognition (SER) presents a significant yet persistent challenge in human-computer interaction. While deep learning has advanced spoken language processing, achieving high performance on limited datasets remains a critical hurdle. This paper confronts this issue by developing and evaluating a suite of machine learning models, including Support Vector Machines (SVMs), Long Short-Term Memory networks (LSTMs), and Convolutional Neural Networks (CNNs), for automated emotion classification in human speech. We demonstrate that by strategically employing transfer learning and innovative data augmentation techniques, our models can achieve impressive performance despite the constraints of a relatively small dataset. Our most effective model, a ResNet34 architecture, establishes a new performance benchmark on the combined RAVDESS and SAVEE datasets, attaining an accuracy of 66.7% and an F1 score of 0.631. These results underscore the substantial benefits of leveraging pre-trained models and data augmentation to overcome data scarcity, thereby paving the way for more robust and generalizable SER systems.
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。
