Limits...
Clutter Mitigation in Echocardiography Using Sparse Signal Separation.

Turek JS, Elad M, Yavneh I - Int J Biomed Imaging (2015)

Bottom Line: In our work, an adaptive approach is used for learning the dictionary from the echo data.MCA is compared to Singular Value Filtering (SVF), a Principal Component Analysis- (PCA-) based filtering technique, and to a high-pass Finite Impulse Response (FIR) filter.Each filter is applied to a simulated hypoechoic lesion sequence, as well as experimental cardiac ultrasound data.

View Article: PubMed Central - PubMed

Affiliation: Department of Computer Science, Israel Institute of Technology (Technion), 3200003 Haifa, Israel.

ABSTRACT
In ultrasound imaging, clutter artifacts degrade images and may cause inaccurate diagnosis. In this paper, we apply a method called Morphological Component Analysis (MCA) for sparse signal separation with the objective of reducing such clutter artifacts. The MCA approach assumes that the two signals in the additive mix have each a sparse representation under some dictionary of atoms (a matrix), and separation is achieved by finding these sparse representations. In our work, an adaptive approach is used for learning the dictionary from the echo data. MCA is compared to Singular Value Filtering (SVF), a Principal Component Analysis- (PCA-) based filtering technique, and to a high-pass Finite Impulse Response (FIR) filter. Each filter is applied to a simulated hypoechoic lesion sequence, as well as experimental cardiac ultrasound data. MCA is demonstrated in both cases to outperform the FIR filter and obtain results comparable to the SVF method in terms of contrast-to-noise ratio (CNR). Furthermore, MCA shows a lower impact on tissue sections while removing the clutter artifacts. In experimental heart data, MCA obtains in our experiments clutter mitigation with an average CNR improvement of 1.33 dB.

No MeSH data available.


Related in: MedlinePlus

Region of interest used to compute CNR as in (12) for the simulated hypoechoic lesion. The white box corresponds to the ROI inside the lesion with clutter artifacts and the black boxes indicate the ROI outside the lesion.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4495184&req=5

fig3: Region of interest used to compute CNR as in (12) for the simulated hypoechoic lesion. The white box corresponds to the ROI inside the lesion with clutter artifacts and the black boxes indicate the ROI outside the lesion.

Mentions: The resulting performances of the algorithms were measured using contrast-to-noise ratio (CNR). The CNR is defined as(12)CNR=20 log10⁡μi−μoσo,where μi and μo are the mean envelope-detected quantities in regions with clutter artifact and without artifacts, respectively, and σo is the standard deviation in the clutter-empty region. Figure 3 shows the regions of interest for computing CNR, the middle box indicating the region with clutter artifacts inside the hypoechoic lesion and the outer boxes indicating the regions outside the lesion. The CNR performance measure may be misleading in some cases. For example, if high values of tissue speckles in the region without artifacts are being reduced, the standard deviation σo may decrease faster than the mean μo and the ratio /μo//σo may increase, making the CNR higher. Subsequently, the CNR may exhibit better values than perfect filtered images. Therefore, performance was measured also using peak signal-to-noise ratio (PSNR) in a few particular cases in order to show further differences that exist between the tested methods. PSNR is computed using the following expression:(13)PSNR=20 log10⁡MAX1/ns−s^22,where MAX is the maximum pixel value in the clean envelope-detected image, s ∈ ℝn is the clutter-free envelope-detected signal, and is the reconstructed envelope-detected signal. PSNR measures the distance to the perfect image, penalizing for any difference from it. Thus, it is capable of measuring the remaining clutter as well as the amount of tissue removed. In contrast to CNR, PSNR requires the perfect filtered signal.


Clutter Mitigation in Echocardiography Using Sparse Signal Separation.

Turek JS, Elad M, Yavneh I - Int J Biomed Imaging (2015)

Region of interest used to compute CNR as in (12) for the simulated hypoechoic lesion. The white box corresponds to the ROI inside the lesion with clutter artifacts and the black boxes indicate the ROI outside the lesion.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4495184&req=5

fig3: Region of interest used to compute CNR as in (12) for the simulated hypoechoic lesion. The white box corresponds to the ROI inside the lesion with clutter artifacts and the black boxes indicate the ROI outside the lesion.
Mentions: The resulting performances of the algorithms were measured using contrast-to-noise ratio (CNR). The CNR is defined as(12)CNR=20 log10⁡μi−μoσo,where μi and μo are the mean envelope-detected quantities in regions with clutter artifact and without artifacts, respectively, and σo is the standard deviation in the clutter-empty region. Figure 3 shows the regions of interest for computing CNR, the middle box indicating the region with clutter artifacts inside the hypoechoic lesion and the outer boxes indicating the regions outside the lesion. The CNR performance measure may be misleading in some cases. For example, if high values of tissue speckles in the region without artifacts are being reduced, the standard deviation σo may decrease faster than the mean μo and the ratio /μo//σo may increase, making the CNR higher. Subsequently, the CNR may exhibit better values than perfect filtered images. Therefore, performance was measured also using peak signal-to-noise ratio (PSNR) in a few particular cases in order to show further differences that exist between the tested methods. PSNR is computed using the following expression:(13)PSNR=20 log10⁡MAX1/ns−s^22,where MAX is the maximum pixel value in the clean envelope-detected image, s ∈ ℝn is the clutter-free envelope-detected signal, and is the reconstructed envelope-detected signal. PSNR measures the distance to the perfect image, penalizing for any difference from it. Thus, it is capable of measuring the remaining clutter as well as the amount of tissue removed. In contrast to CNR, PSNR requires the perfect filtered signal.

Bottom Line: In our work, an adaptive approach is used for learning the dictionary from the echo data.MCA is compared to Singular Value Filtering (SVF), a Principal Component Analysis- (PCA-) based filtering technique, and to a high-pass Finite Impulse Response (FIR) filter.Each filter is applied to a simulated hypoechoic lesion sequence, as well as experimental cardiac ultrasound data.

View Article: PubMed Central - PubMed

Affiliation: Department of Computer Science, Israel Institute of Technology (Technion), 3200003 Haifa, Israel.

ABSTRACT
In ultrasound imaging, clutter artifacts degrade images and may cause inaccurate diagnosis. In this paper, we apply a method called Morphological Component Analysis (MCA) for sparse signal separation with the objective of reducing such clutter artifacts. The MCA approach assumes that the two signals in the additive mix have each a sparse representation under some dictionary of atoms (a matrix), and separation is achieved by finding these sparse representations. In our work, an adaptive approach is used for learning the dictionary from the echo data. MCA is compared to Singular Value Filtering (SVF), a Principal Component Analysis- (PCA-) based filtering technique, and to a high-pass Finite Impulse Response (FIR) filter. Each filter is applied to a simulated hypoechoic lesion sequence, as well as experimental cardiac ultrasound data. MCA is demonstrated in both cases to outperform the FIR filter and obtain results comparable to the SVF method in terms of contrast-to-noise ratio (CNR). Furthermore, MCA shows a lower impact on tissue sections while removing the clutter artifacts. In experimental heart data, MCA obtains in our experiments clutter mitigation with an average CNR improvement of 1.33 dB.

No MeSH data available.


Related in: MedlinePlus