This paper proposes a personalized head-related transfer function (HRTF) estimation method based on deep neural networks by using anthropometric measurements and ear images. The proposed method consists of three sub-networks for representing personalized features and estimating the HRTF. As input features for neural networks, the anthropometric measurements regarding the head and torso are used for a feedforward deep neural network (DNN), and the ear images are used for a convolutional neural network (CNN). After that, the outputs of these two sub-networks are merged into another DNN for estimation of the personalized HRTF. To evaluate the performance of the proposed method, objective and subjective evaluations are conducted. For the objective evaluation, the root mean square error (RMSE) and the log spectral distance (LSD) between the reference HRTF and the estimated one are measured. Consequently, the proposed method provides the RMSE of −18.40 dB and LSD of 4.47 dB, which are lower by 0.02 dB and higher by 0.85 dB than the DNN-based method using anthropometric data without pinna measurements, respectively. Next, a sound localization test is performed for the subjective evaluation. As a result, it is shown that the proposed method can localize sound sources with higher accuracy of around 11% and 6% than the average HRTF method and DNN-based method, respectively. In addition, the reductions of the front/back confusion rate by 12.5% and 2.5% are achieved by the proposed method, compared to the average HRTF method and DNN-based method, respectively.

Autism spectrum disorder (ASD) is a developmental disorder with a life-span disability. While diagnostic instruments have been developed and qualified based on the accuracy of the discrimination of children with ASD from typical development (TD) children, the stability of such procedures can be disrupted by limitations pertaining to time expenses and the subjectivity of clinicians. Consequently, automated diagnostic methods have been developed for acquiring objective measures of autism, and in various fields of research, vocal characteristics have not only been reported as distinctive characteristics by clinicians, but have also shown promising performance in several studies utilizing deep learning models based on the automated discrimination of children with ASD from children with TD. However, difficulties still exist in terms of the characteristics of the data, the complexity of the analysis, and the lack of arranged data caused by the low accessibility for diagnosis and the need to secure anonymity. In order to address these issues, we introduce a pre-trained feature extraction auto-encoder model and a joint optimization scheme, which can achieve robustness for widely distributed and unrefined data using a deep-learning-based method for the detection of autism that utilizes various models. By adopting this auto-encoder-based feature extraction and joint optimization in the extended version of the Geneva minimalistic acoustic parameter set (eGeMAPS) speech feature data set, we acquire improved performance in the detection of ASD in infants compared to the raw data set.

In this paper, a multi-task learning U-shaped neural network (MTU-Net) is proposed and applied to single-channel speech enhancement (SE). The proposed MTU-based SE method estimates an ideal binary mask (IBM) or an ideal ratio mask (IRM) by extending the decoding network of a conventional U-Net to simultaneously model the speech and noise spectra as the target. The effectiveness of the proposed SE method was evaluated under both matched and mismatched noise conditions between training and testing by measuring the perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI). Consequently, the proposed SE method with IRM achieved a substantial improvement with higher average PESQ scores by 0.17, 0.52, and 0.40 than other state-of-the-art deep-learning-based methods, such as the deep recurrent neural network (DRNN), SE generative adversarial network (SEGAN), and conventional U-Net, respectively. In addition, the STOI scores of the proposed SE method are 0.07, 0.05, and 0.05 higher than those of the DRNN, SEGAN, and U-Net, respectively. Next, voice activity detection (VAD) is also proposed by using the IRM estimated by the proposed MTU-Net-based SE method, which is fundamentally an unsupervised method without any model training. Then, the performance of the proposed VAD method was compared with the performance of supervised learning-based methods using a deep neural network (DNN), a boosted DNN, and a long short-term memory (LSTM) network. Consequently, the proposed VAD methods show a slightly better performance than the three neural network-based methods under mismatched noise conditions.

In this paper, we propose an end-to-end (E2E) neural network model to detect autism spectrum disorder (ASD) from children’s voices without explicitly extracting the deterministic features. In order to obtain the decisions for discriminating between the voices of children with ASD and those with typical development (TD), we combined two different feature-extraction models and a bidirectional long short-term memory (BLSTM)-based classifier to obtain the ASD/TD classification in the form of probability. We realized one of the feature extractors as the bottleneck feature from an autoencoder using the extended version of the Geneva minimalistic acoustic parameter set (eGeMAPS) input. The other feature extractor is the context vector from a pretrained wav2vec2.0-based model directly applied to the waveform input. In addition, we optimized the E2E models in two different ways: (1) fine-tuning and (2) joint optimization. To evaluate the performance of the proposed E2E models, we prepared two datasets from video recordings of ASD diagnoses collected between 2016 and 2018 at Seoul National University Bundang Hospital (SNUBH), and between 2019 and 2021 at a Living Lab. According to the experimental results, the proposed wav2vec2.0-based E2E model with joint optimization achieved significant improvements in the accuracy and unweighted average recall, from 64.74% to 71.66% and from 65.04% to 70.81%, respectively, compared with a conventional model using autoencoder-based BLSTM and the deterministic features of the eGeMAPS.

In this paper, a new two-step joint optimization approach based on the asynchronous subregion optimization method is proposed for training a pipeline model composed of two different models. The first-step processing of the proposed joint optimization approach trains the front-end model only, and the second-step processing trains all the parameters of the combined model together. In the asynchronous subregion optimization method, the first-step processing only supports the goal of the front-end model. However, the first-step processing of the proposed approach works with a new loss function to make the front-end model support the goal of the back-end model. The proposed optimization approach was applied, here, to a pipeline composed of a deep complex convolutional recurrent network (DCCRN)-based speech enhancement model and a conformer-transducer-based ASR model as a front-end and a back-end, respectively. Then, the performance of the proposed two-step joint optimization approach was evaluated on the LibriSpeech automatic speech recognition (ASR) corpus in noisy environments by measuring the character error rate (CER) and word error rate (WER). In addition, an ablation study was carried out to examine the effectiveness of the proposed optimization approach on each of the processing blocks in the conformer-transducer ASR model. Consequently, it was shown from the ablation study that the conformer-transducer-based ASR model with the joint network trained only by the proposed optimization approach achieved the lowest average CER and WER. Moreover, the proposed optimization approach reduced the average CER and WER on the Test-Noisy dataset under matched noise conditions by 0.30% and 0.48%, respectively, compared to the approach of separate optimization of speech enhancement and ASR. Compared to the conventional two-step joint optimization approach, the proposed optimization approach provided average CER and WER reductions of 0.22% and 0.31%, respectively. Moreover, it was revealed that the proposed optimization approach achieved a lower average CER and WER, by 0.32% and 0.43%, respectively, than the conventional optimization approach under mismatched noise conditions.

This paper addresses a joint training approach applied to a pipeline comprising speech enhancement (SE) and automatic speech recognition (ASR) models, where an acoustic tokenizer is included in the pipeline to leverage the linguistic information from the ASR model to the SE model. The acoustic tokenizer takes the outputs of the ASR encoder and provides a pseudo-label through K-means clustering. To transfer the linguistic information, represented by pseudo-labels, from the acoustic tokenizer to the SE model, a cluster-based pairwise contrastive (CBPC) loss function is proposed, which is a self-supervised contrastive loss function, and combined with an information noise contrastive estimation (infoNCE) loss function. This combined loss function prevents the SE model from overfitting to outlier samples and represents the pronunciation variability in samples with the same pseudo-label. The effectiveness of the proposed CBPC loss function is evaluated on a noisy LibriSpeech dataset by measuring both the speech quality scores and the word error rate (WER). The experimental results reveal that the proposed joint training approach using the described CBPC loss function achieves a lower WER than the conventional joint training approaches. In addition, it is demonstrated that the speech quality scores of the SE model trained using the proposed training approach are higher than those of the standalone-SE model and SE models trained using conventional joint training approaches. An ablation study is also conducted to investigate the effects of different combinations of loss functions on the speech quality scores and WER. Here, it is revealed that the proposed CBPC loss function combined with infoNCE contributes to a reduced WER and an increase in most of the speech quality scores.