Petr Schaffer

Bio: Petr Schaffer is an academic researcher from Dresden University of Technology. The author has contributed to research in topics: Speech recognition & Computer science. The author has an hindex of 2, co-authored 3 publications receiving 65 citations.

Silent speech command word recognition using stepped frequency continuous wave radar

Abstract: Recovering speech in the absence of the acoustic speech signal itself, i.e., silent speech, holds great potential for restoring or enhancing oral communication in those who lost it. Radar is a relatively unexplored silent speech sensing modality, even though it has the advantage of being fully non-invasive. We therefore built a custom stepped frequency continuous wave radar hardware to measure the changes in the transmission spectra during speech between three antennas, located on both cheeks and the chin with a measurement update rate of 100 Hz. We then recorded a command word corpus of 40 phonetically balanced, two-syllable German words and the German digits zero to nine for two individual speakers and evaluated both the speaker-dependent multi-session and inter-session recognition accuracies on this 50-word corpus using a bidirectional long-short term memory network. We obtained recognition accuracies of 99.17% and 88.87% for the speaker-dependent multi-session and inter-session accuracy, respectively. These results show that the transmission spectra are very well suited to discriminate individual words from one another, even across different sessions, which is one of the key challenges for fully non-invasive silent speech interfaces.

A user-friendly headset for radar-based silent speech recognition

Abstract: Silent speech interfaces allow speech communication to take place in the absence of the acoustic speech signal. Radar-based sensing with radio antennas on the speakers’ face can be used as a non-invasive modality to measure speech articulation in such applications. One of the major challenges with this approach is the variability between different sessions, mainly due to the repositioning of the antennas on the face of the speaker. In order to reduce the impact of this influencing factor, we developed a wearable headset that can be 3D-printed with flexible materials and weighs only about 69 g. For evaluation, a radar-based word recognition experiment was performed, where five speakers recorded a speech corpus in multiple sessions, alternatively with the headset and with double-sided tape to place the antennas on the face. By using a bidirectional long short-term memory network for classification, an average inter-session word accuracy of 76.50% and 68.18% was obtained using the headset and the tape, respectively. This indicates that the antenna (re-) positioning accuracy with the headset is not worse than that with the double-sided tape while providing other benefits.

On the Optimal Set of Features and the Robustness of Classifiers in Radar-based Silent Phoneme Recognition

Abstract: Silent speech recognition (SSR) is an active area of research with applications ranging from speech restoration to speech enhancement. Radar-based SSR has been proposed and investigated as a non-invasive method to infer vocal tract states and articulatory movements from measured changes in scattering parameters. One of the challenges in developing a radar-based SSR system is to determine the optimal set of features from these measurements. In this study, we therefore investigated the following problems: (a) The selection of the features that play the most significant role for classification. (b) The determination of the contribution of each reflection and transmission spectrum and the most important frequencies. (c) The determination of the performance of the classifiers when using fewer features. (d) The determination of the robustness of the classifiers against different noise levels. The data used in this study consisted of 230 samples of 25 German phonemes (15 vowels, each in 10 contexts, and 10 consonants, each in 8 contexts) produced by two German native speakers. Using the full feature set, a Linear Discriminant Analysis (LDA) classifier achieved up to 94 % classification accuracy for speaker 1 and 84 % for speaker 2. Using only the most important features as identified by a decision tree, the classification accuracy deteriorated slightly in most conditions, but in one case improved the accuracy from 73.5 % to 81 %. Regarding the robustness against noise, the accuracy of the LDA dropped sharply with increasing noise levels, while the decrease of the SVM’s accuracy was less steep.

Monthly Notices of the Royal Astronomical Society

Abstract: Monthly Notices is one of the three largest general primary astronomical research publications. It is an international journal, published by the Royal Astronomical Society. This article 1 describes its publication policy and practice.

Dust-to-Gas and Refractory-to-Ice Mass Ratios of Comet 67P/Churyumov-Gerasimenko from Rosetta Observations

Abstract: This chapter reviews the estimates of the dust-to-gas and refractory-to-ice mass ratios derived from Rosetta measurements in the lost materials and the nucleus of 67P/Churyumov-Gerasimenko, respectively. First, the measurements by Rosetta instruments are described, as well as relevant characteristics of 67P. The complex picture of the activity of 67P, with its extreme North-South seasonal asymmetry, is presented. Individual estimates of the dust-to-gas and refractory-to-ice mass ratios are then presented and compared, showing wide ranges of plausible values. Rosetta’s wealth of information suggests that estimates of the dust-to-gas mass ratio made in cometary comae at a single point in time may not be fully representative of the refractory-to-ice mass ratio within the cometary nuclei being observed.

Observations, Meteorites, and Models: A Preflight Assessment of the Composition and Formation of (16) Psyche.

Abstract: Some years ago, the consensus was that asteroid (16) Psyche was almost entirely metal. New data on density, radar properties, and spectral signatures indicate that the asteroid is something perhaps even more enigmatic: a mixed metal and silicate world. Here we combine observations of Psyche with data from meteorites and models for planetesimal formation to produce the best current hypotheses for Psyche's properties and provenance. Psyche's bulk density appears to be between 3,400 and 4,100 kg m-3. Psyche is thus predicted to have between ~30 and ~60 vol% metal, with the remainder likely low-iron silicate rock and not more than ~20% porosity. Though their density is similar, mesosiderites are an unlikely analog to bulk Psyche because mesosiderites have far more iron-rich silicates than Psyche appears to have. CB chondrites match both Psyche's density and spectral properties, as can some pallasites, although typical pallasitic olivine contains too much iron to be consistent with the reflectance spectra. Final answers, as well as resolution of contradictions in the data set of Psyche physical properties, for example, the thermal inertia measurements, may not be resolved until the NASA Psyche mission arrives in orbit at the asteroid. Despite the range of compositions and formation processes for Psyche allowed by the current data, the science payload of the Psyche mission (magnetometers, multispectral imagers, neutron spectrometer, and a gamma-ray spectrometer) will produce data sets that distinguish among the models.