Share This Author
Automatic detection of mild cognitive impairment from spontaneous speech using ASR
This work automates the extraction of the features of Mild Cognitive Impairment by applying automatic speech recognition (ASR), and uses machine learning methods to separate the subjects with MCI from the control group.
Identifying Mild Cognitive Impairment and mild Alzheimer's disease based on spontaneous speech using ASR and linguistic features
A Speech Recognition-based Solution for the Automatic Detection of Mild Cognitive Impairment from Spontaneous Speech
The temporal analysis of spontaneous speech can be exploited in implementing a new, auto-matic detection-based tool for screening MCI for the community.
DNN-Based Ultrasound-to-Speech Conversion for a Silent Speech Interface
- T. Csapó, Tamás Grósz, G. Gosztolya, L. Tóth, Alexandra Markó
- Computer ScienceINTERSPEECH
- 20 August 2017
It is found that the representation that used several neighboring image frames in combination with a feature selection method was preferred both by the subjects taking part in the listening experiments, and in terms of the Normalized Mean Squared Error.
Cross-lingual portability of MLP-based tandem features - a case study for English and Hungarian
This work examines the portability of feature extractor MLPs between an Indo-European and a Finno-Ugric language and finds that the cross-lingual configurations achieve similar performance to the monolingual system, and that the AF detectors lead to slightly worse performance, despite the expectation that they should be more language-independent than phone-based MLPs.
On evaluation metrics for social signal detection
- G. Gosztolya
- Computer ScienceINTERSPEECH
It is argued that the Area Under Curve metric is not really suitable for social signals detection, and it is shown that applying a very simple smoothing function on the output of the framelevel scores of state-of-the-art classifiers can significantly improve the AUC scores, but perform poorly when employed in a Hidden Markov Model.
User-centric Evaluation of Automatic Punctuation in ASR Closed Captioning
A user-centric evaluation of a real-time closed captioning system enhanced by a lightweight RNN-based punctuation module confirms that automatic punctuation itself significantly increases understandability, even if several other factors interplay in subjective impression.
Ultrasound-based Articulatory-to-Acoustic Mapping with WaveGlow Speech Synthesis
- Tam'as G'abor Csap'o, Csaba Zaink'o, L. T'oth, G. Gosztolya, Alexandra Mark'o
- 6 August 2020
This paper uses a flow-based neural vocoder (WaveGlow) pre-trained on a large amount of English and Hungarian speech data to train a convolutional neural network for articulatory-to-acoustic mapping using deep neural networks.
Assessing the degree of nativeness and parkinson's condition using Gaussian processes and deep rectifier neural networks
Both DNN and GPR are competitive with the baseline SVM, and that the results can be improved further by combining the classifiers, but by far the best results are obtained when a speaker clustering method is applied to identify the files that belong to the same speaker.
Detecting autism, emotions and social signals using adaboost
This paper treats sub-challenges of paralinguistic detection, categorizing whole (albeit short) recordings by speaker emotion, conflict or the presence of development disorders (autism) as general classification tasks and applies the general-purpose machine learning meta-algorithm, AdaBoost.MH, and its recently proposed variant, Ada boost.BA, to them.