The Generalized Contrast-to-Noise Ratio: A Formal Definition for Lesion Detectability

Rodríguez-Molares, Alfonso; Rindal, Ole Marius Hoel; D'hooge, Jan; Måsøy, Svein-Erik; Austeng, Andreas; Bell, Muyinatu A. Lediju; Torp, Hans

doi:10.1109/tuffc.2019.2956855

Cited by 302 publications

(148 citation statements)

References 83 publications

Supporting

Mentioning

145

Contrasting

Unclassified

Order By: Relevance

“…The generalized contrast-to-noise ratio (gCNR) 62 – 64 was used to assess the likelihood of discrimination between regions of interest (ROIs) of beamformed photoacoustic data and after normalization but before the log compression stage (i.e., the first optimization criterion for parameter selection).

where

and

are the probability density functions of signal amplitudes within ROIs inside and outside the target, respectively.…”

Section: Methodsmentioning

confidence: 99%

GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions

González

Bell

2020

J. Biomed. Opt.

View full text Add to dashboard Cite

Significance: Photoacoustic-based visual servoing is a promising technique for surgical tool tip tracking and automated visualization of photoacoustic targets during interventional procedures. However, one outstanding challenge has been the reliability of obtaining segmentations using low-energy light sources that operate within existing laser safety limits. Aim: We developed the first known graphical processing unit (GPU)-based real-time implementation of short-lag spatial coherence (SLSC) beamforming for photoacoustic imaging and applied this real-time algorithm to improve signal segmentation during photoacoustic-based visual servoing with low-energy lasers. Approach: A 1-mm-core-diameter optical fiber was inserted into ex vivo bovine tissue. Photoacoustic-based visual servoing was implemented as the fiber was manually displaced by a translation stage, which provided ground truth measurements of the fiber displacement. GPU-SLSC results were compared with a central processing unit (CPU)-SLSC approach and an amplitude-based delay-and-sum (DAS) beamforming approach. Performance was additionally evaluated with in vivo cardiac data. Results: The GPU-SLSC implementation achieved frame rates up to 41.2 Hz, representing a factor of 348 speedup when compared with offline CPU-SLSC. In addition, GPU-SLSC successfully recovered low-energy signals (i.e., ≤268 μJ) with mean ± standard deviation of signal-to-noise ratios of 11.2 AE 2.4 (compared with 3.5 AE 0.8 with conventional DAS beamforming). When energies were lower than the safety limit for skin (i.e., 394.6 μJ for 900-nm wavelength laser light), the median and interquartile range (IQR) of visual servoing tracking errors obtained with GPU-SLSC were 0.64 and 0.52 mm, respectively (which were lower than the median and IQR obtained with DAS by 1.39 and 8.45 mm, respectively). GPU-SLSC additionally reduced the percentage of failed segmentations when applied to in vivo cardiac data. Conclusions: Results are promising for the use of low-energy, miniaturized lasers to perform GPU-SLSC photoacoustic-based visual servoing in the operating room with laser pulse repetition frequencies as high as 41.2 Hz.

show abstract

where

and

are the probability density functions of signal amplitudes within ROIs inside and outside the target, respectively.…”

Section: Methodsmentioning

confidence: 99%

GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions

González

Bell

2020

J. Biomed. Opt.

View full text Add to dashboard Cite

show abstract

“…The choice to automatically use a small circular region about the cyst center was made to avoid manual ROI selection across the thousands of simulation and phantom test sets, yet still ensure that the results would be a meaningful assessment of the difference in signal amplitude inside and outside the detected cyst region. This automated ROI selection additionally is intended to prevent the inclusion of misclassifications (e.g., cyst pixels at the cyst boundary detectected as tissue and vice versa), which are instead evaluated with the gCNR metric [51]. The ROI outside of the cyst was the same size as the inside ROI and was located at the same depth as the cyst.…”

Section: J Evaluation Metricsmentioning

confidence: 99%

“…The enhanced beamformed image contains the same tissue background as the DAS beamformed image and therefore has identical SNR to the DAS beamformed image. 4) Generalized Contrast-to-Noise Ratio (gCNR): The gCNR was recently introduced as a more accurate measure of lesion detectability in comparison to CNR [51], and it calculated as:…”

Section: J Evaluation Metricsmentioning

confidence: 99%

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Nair

Washington

Tran

et al. 2020

IEEE Trans. Ultrason., Ferroelect., Freq. Contr.

Self Cite

View full text Add to dashboard Cite

Single plane wave transmissions are promising for automated imaging tasks requiring high ultrasound frame rates over an extended field of view. However, a single plane wave insonification typically produces suboptimal image quality. To address this limitation, we are exploring the use of deep neural networks (DNNs) as an alternative to delay-and-sum beamforming. The objectives of this work are to obtain information directly from raw channel data and to simultaneously generate both a segmentation map for automated ultrasound tasks and a corresponding ultrasound B-mode image for interpretable supervision of the automation. We focus on visualizing and segmenting anechoic targets surrounded by tissue and ignoring or de-emphasizing less important surrounding structures. DNNs trained with Field II simulations were tested with simulated, experimental phantom, and in vivo datasets that were not included during training. With unfocused input channel data (i.e., prior to the application of receive time delays), simulated, experimental phantom, and in vivo test datasets achieved mean ± standard deviation Dice similarity coefficients of 0.92 ± 0.13, 0.92±0.03, and 0.77±0.07, respectively, and generalized contrastto-noise ratios (gCNR) of 0.95±0.08, 0.93±0.08, and 0.75±0.14, respectively. With subaperture beamformed channel data and a modification to the input layer of the DNN architecture to accept these data, the fidelity of image reconstruction increased (e.g., mean gCNR of multiple acquisitions of two in vivo breast cysts ranged 0.89-0.96), but DNN display frame rates were reduced from 395 Hz to 287 Hz. Overall, the DNNs successfully translated feature representations learned from simulated data to phantom and in vivo data, which is promising for this novel approach to simultaneous ultrasound image formation and segmentation.

show abstract

“…In addition, the generalized contrast-to-noise ratio (gCNR) [41] was also assessed to evaluate the imaging performances of different methods. The gCNR compares the overlap between the intensity distributions of background region and cyst region, and it is defined as…”

Section: Image Quality Metricsmentioning

confidence: 99%

Neighborhood Singular Value Decomposition Filter and Application in Adaptive Beamforming for Coherent Plane-Wave Compounding

et al. 2020

View full text Add to dashboard Cite

Coherent plane-wave compounding (CPWC) is widely used in medical ultrasound imaging, in which plane-waves tilted at multiple angles are used to reconstruct ultrasound images. CPWC helps to achieve a balance between frame rate and image quality. However, the image quality of CPWC is limited due to sidelobes and noise interferences. Filtering techniques and adaptive beamforming methods are commonly used to suppress noise and sidelobes. Here, we propose a neighborhood singular value decomposition (NSVD) filter to obtain high-quality images in CPWC. The NSVD filter is applied to adaptive beamforming by combining with adaptive weighting factors. The NSVD filter is advantageous because of its singular value decomposition (SVD) and smoothing filters, performing the SVD processing in neighboring regions while using a sliding rectangular window to filter the entire imaging region. We also tested the application of NSVD in adaptive beamforming. The NSVD filter was combined with short-lag spatial coherence (SLSC), coherence factor (CF), and generalized coherence factor (GCF) to enhance performances of adaptive beamforming methods. The proposed methods were evaluated using simulated and experimental datasets. We found that NSVD can suppress noise and achieve improved contrast (contrast ratio (CR), contrast-to-noise ratio (CNR) and generalized CNR (gCNR)) compared to CPWC. When the NSVD filter is used, adaptive weighting methods provide higher CR, CNR, gCNR and speckle signal-to-noise ratio (sSNR), indicating that NSVD is able to improve the imaging performance of adaptive beamforming in noise suppression and speckle pattern preservation.

show abstract

The Generalized Contrast-to-Noise Ratio: A Formal Definition for Lesion Detectability

Cited by 302 publications

References 83 publications

GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions

GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Neighborhood Singular Value Decomposition Filter and Application in Adaptive Beamforming for Coherent Plane-Wave Compounding

Contact Info

Product

Resources

About