Alexander Schasse

Estimation of subband speech correlations for noise reduction via MVDR Processing

Recently, it has been proposed to use the minimum-variance distortionless-response (MVDR) approach in single-channel speech enhancement in the short-time frequency domain. By applying optimal FIR filters to each subband signal, these filters reduce additive noise components with less speech distortion compared to conventional approaches. An important ingredient to these filters is the temporal correlation of the speech signals. We derive algorithms to provide a blind estimation of this quantity based on a maximum-likelihood and maximum a-posteriori estimation. To derive proper models for the inter-frame correlation of the speech and noise signals, we investigate their statistics on a large dataset. If the speech correlation is properly estimated, the previously derived subband filters discussed in this work show significantly less speech distortion compared to conventional noise reduction algorithms. Therefore, the focus of the experimental parts of this work lies on the quality and intelligibility of the processed signals. To evaluate the performance of the subband filters in combination with the clean speech inter-frame correlation estimators, we predict the speech quality and intelligibility by objective measures.

An Automated System for Full Angle Spatial Compounding in Ultrasound Breast Imaging

For the detection of breast cancer, ultrasound is conventionally used in addition to mammography. However, ultrasound is highly operator-dependent since speckle, depth dependency and artifacts (e.g. shadowing) affect image quality. Full Angle Spatial Compounding (FASC) may overcome those limitations by superimposing ultrasound data acquired in one cross-sectional plane from aspect angles all around an object. Furthermore, we showed in vitro that FASC also improves contrast perfusion imaging. To apply FASC to breast imaging, we developed an add-on system to a conventional ultrasound scanner, integrated in a custom-made examination couch. Here, we present first in-vivo results from test persons and patients with breast lesions.

Ultrasound breast imaging using Full Angle Spatial Compounding: In-vivo results

For the detection of breast cancer, ultrasound is conventionally used in addition to mammography. However, ultrasound is highly operator-dependent since speckle, depth dependency and other artifacts (e.g. shadowing) affect image quality. Full Angle Spatial Compounding (FASC) may overcome those limitations by superimposing ultrasound data acquired in one cross-sectional plane from aspect angles all around an object. Furthermore, we showed in vitro that FASC also improves contrast perfusion imaging. To apply FASC to breast imaging, we developed an add-on system to a conventional ultrasound scanner, integrated in a custom-made examination couch. Here, we present first in-vivo results from test persons and patients with breast lesions.

9C-2 Reconstruction of Speed of Sound for a Correction of Transit Time in Full Angle Spatial Compounding

For full angle spatial compounding, an object is imaged in one plane from multiple aspect angles and the obtained images are superimposed in correct geometric orientation. Thus, information about the distribution of speed of sound inside the object is required for a correction of individual images with respect to transit time and refraction of the ultrasonic waves. In this paper, we present a method to reconstruct such a distribution by a filtered backprojection of echo data from a reflector positioned behind the object. Using the reconstructed distribution, individual ultrasound images were corrected by a ray tracing algorithm before spatial compounding. Here, refraction was only accounted for at the outer boundary of the object, while inside only transit time was corrected. In vitro analyses are presented that prove the applicability of this approach.

Online inter-frame correlation estimation methods for speech enhancement in frequency subbands

In this paper, we propose solutions for the online adaptation of optimal FIR filters for speech enhancement in DFT subbands. An important ingredient to such filters is the estimation of the inter-frame correlation of the clean speech signal. While this correlation was assumed to be perfectly known in former studies, we discuss two online estimation approaches based on a constant noise inter-frame correlation and on the use of a binary mask. We show that a filtering of subband signals based on these estimated quantities outperforms a conventional, instantaneous spectral weighting, such as the frequency-domain Wiener filter at least for high SNR conditions.

Ultrasound Computed Tomography in Breast Imaging: First Clinical Results of a Custom-Made Scanner

PURPOSE To test a system using ultrasound computed tomography (USCT) that superimposes ultrasound data acquired in one cross-sectional plane from multiple angles around the breast (Full Angle Spatial Compounding, FASC) and to reconstruct the distribution of the speed of sound in the breast (SoS reconstruction). MATERIALS AND METHODS We developed a system combining a conventional ultrasound scanner with a PC-controlled mechanical setup integrated in a custom-made examination couch. In a feasibility study, 3 volunteers (age 26 - 74 years) and one patient with breast cancer were studied. Subjects were placed in the prone position on this couch, with the breast hanging in a water tank. The ultrasound probe was moved in several planes around the breast. A curved reflector that followed the movement of the probe behind the breast was used to calculate the SoS within the breast tissue. Echo-data was processed offline by custom-made software to calculate both FASC and SoS images. RESULTS In FASC images a reduction of artifacts (i. e. shadowing of Coopers ligaments and irregular edges of inhomogeneous lesions) and speckles as well as clear visualization of the inner architecture of the breast was achieved. SoS images delivered further diagnostic information and helped to compensate for geometric distortions in the computed images. Difficulties in the visualization of lesions near the thoracic wall and/or the axillary are limitations of this technique. CONCLUSION The first clinical results of USCT imaging have proven its feasibility as an automated and standardized technique for breast imaging.

Two-stage filter-bank system for improved single-channel noise reduction in hearing aids

The filter-bank system implemented in hearing aids has to fulfill various constraints such as low latency and high stop-band attenuation, usually at the cost of low frequency resolution. In the context of frequency-domain noise-reduction algorithms, insufficient frequency resolution may lead to annoying residual noise artifacts since the spectral harmonics of the speech cannot properly be resolved. Especially in case of female speech signals, the noise between the spectral harmonics causes a distinct roughness of the processed signals. Therefore, this work proposes a two-stage filter-bank system, such that the frequency resolution can be improved for the purpose of noise reduction, while the original first-stage hearing-aid filter-bank system can still be used for compression and amplification. We also propose methods to implement the second filter-bank stage with little additional algorithmic delay. Furthermore, the computational complexity is an important design criterion. This finally leads to an application of the second filter-bank stage to lower frequency bands only, resulting in the ability to resolve the harmonics of speech. The paper presents a systematic description of the second filter-bank stage, discusses its influence on the processed signals in detail and further presents the results of a listening test which indicates the improved performance compared to the original single-stage filter-bank system.

Efficient estimation of inter-subband speech correlations

We propose an approach to compute the inter-subband correlation (ISBC) of noisy speech signals to distinguish between speech and noise segments in the time-frequency plane. The proposed spectral correlation estimator provides information about the input signal which can be used to derive a binary mask or the speech-presence probability. Unlike other approaches it does not require an estimate of the noise power. To this end we analyse a received noisy speech signal in the modulation domain and identify similarly modulated subband signals within a range of modulation frequencies that are typical for speech signals. Based on this pre-processing step, we identify a single reference subband that most likely contains speech and estimate the spectral correlations with respect to this reference band. The algorithm proposed in this paper aims at a very low computational complexity which makes it suitable for hearing aids.