Deep Cross-Corpus Speech Emotion Recognition: Recent Advances and Perspectives

Automatic speech emotion recognition (SER) is a challenging component of human-computer interaction (HCI). Existing literatures mainly focus on evaluating the SER performance by means of training and testing on a single corpus with a single language setting. However, in many practical applications, there are great differences between the training corpus and testing corpus. Due to the diversity of different speech emotional corpus or languages, most previous SER methods do not perform well when applied in real-world cross-corpus or cross-language scenarios. Inspired by the powerful feature learning ability of recently-emerged deep learning techniques, various advanced deep learning models have increasingly been adopted for cross-corpus SER. This paper aims to provide an up-to-date and comprehensive survey of cross-corpus SER, especially for various deep learning techniques associated with supervised, unsupervised and semi-supervised learning in this area. In addition, this paper also highlights different challenges and opportunities on cross-corpus SER tasks, and points out its future trends.

Improving Cross-Corpus Speech Emotion Recognition with Adversarial Discriminative Domain Generalization (ADDoG)

Automatic speech emotion recognition provides computers with critical context to enable user understanding. While methods trained and tested within the same dataset have been shown successful, they often fail when applied to unseen datasets. To address this, recent work has focused on adversarial methods to find more generalized representations of emotional speech. However, many of these methods have issues converging, and only involve datasets collected in laboratory conditions. In this paper, we introduce Adversarial Discriminative Domain Generalization (ADDoG), which follows an easier to train "meet in the middle" approach. The model iteratively moves representations learned for each dataset closer to one another, improving cross-dataset generalization. We also introduce Multiclass ADDoG, or MADDoG, which is able to extend the proposed method to more than two datasets, simultaneously. Our results show consistent convergence for the introduced methods, with significantly improved results when not using labels from the target dataset. We also show how, in most cases, ADDoG and MADDoG can be used to improve upon baseline state-of-the-art methods when target dataset labels are added and in-the-wild data are considered. Even though our experiments focus on cross-corpus speech emotion, these methods could be used to remove unwanted factors of variation in other settings.

Fusing Visual Attention CNN and Bag of Visual Words for Cross-Corpus Speech Emotion Recognition

Speech emotion recognition (SER) classifies emotions using low-level features or a spectrogram of an utterance. When SER methods are trained and tested using different datasets, they have shown performance reduction. Cross-corpus SER research identifies speech emotion using different corpora and languages. Recent cross-corpus SER research has been conducted to improve generalization. To improve the cross-corpus SER performance, we pretrained the log-mel spectrograms of the source dataset using our designed visual attention convolutional neural network (VACNN), which has a 2D CNN base model with channel- and spatial-wise visual attention modules. To train the target dataset, we extracted the feature vector using a bag of visual words (BOVW) to assist the fine-tuned model. Because visual words represent local features in the image, the BOVW helps VACNN to learn global and local features in the log-mel spectrogram by constructing a frequency histogram of visual words. The proposed method shows an overall accuracy of 83.33%, 86.92%, and 75.00% in the Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS), the Berlin Database of Emotional Speech (EmoDB), and Surrey Audio-Visual Expressed Emotion (SAVEE), respectively. Experimental results on RAVDESS, EmoDB, SAVEE demonstrate improvements of 7.73%, 15.12%, and 2.34% compared to existing state-of-the-art cross-corpus SER approaches.