Mardava Rajugopal Gubbi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mardava Rajugopal Gubbi has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Muyinatu A. Lediju Bell
Department of Electrical and Computer Engineering, Johns Hopkins University, Baltimore, USA
7 shared publications - Eduardo A. Gonzalez
Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, USA
3 shared publications Jonathan Chrispin
Johns Hopkins Medical Institutions
2 shared publications- Alycen Wiacek
Department of Electrical and Computer Engineering Oakland University Rochester Michigan USA
1 shared publication Aravindan Kolandaivelu
Johns Hopkins Medical Institutions
1 shared publication- Derek Allman
Department of Electrical and Computer Engineering, Johns Hopkins University, Baltimore, MD, USA
1 shared publication
Latest Publications
- Top Cited PapersOpen Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 73 (1) , 417-431
In Vivo Demonstration of Deep Learning-Based Photoacoustic Visual Servoing System
Objective: To develop the first known deep learning-based photoacoustic visual servoing system utilizing point source localization and hybrid position-force control to track catheter tips in three dimensions in real-time. Methods: We integrated either object detection or instance segmentation-based localization with hybrid position-force control to create our novel system. Cardiac catheter tips were then tracked across distances of 40 mm in a plastisol phantom and 25-64 mm in an in vivo swine in real-time in nine visual servoing trials total. Results: Object detection-based localization identified the cardiac catheter tip in 88.0-91.7% and 66.7-70.4% of phantom and in vivo channel data frames, respectively. Instance segmentation detection rates ranged 86.4-100.0% in vivo. These catheter tips were tracked with errors as low as 0.5 mm in phantom trials and 0.8 mm in the in vivo trials. The mean inference times were ≥145.3 ms and ≥516.3 ms with object detection-based and instance segmentation-based point source localization, respectively. Hybrid position-force control system enabled contact with the imaging surface during ≥99.43% of each visual servoing trial. Conclusion: Our novel deep learning-based photoacoustic visual servoing system was successfully demonstrated. Object detectionbased localization operated with inference times that are more suitable for real-time implementations while instance segmentation had lower tracking errors. Significance: After implementing suggested optimization modifications, our novel system has the potential to track catheter tips, needle tips, and other surgical tool tips in real-time during surgical and interventional procedures.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 73 (1) , 417-431
Top Cited PapersOpen AccessIn Vivo Demonstration of Deep Learning-Based Photoacoustic Visual Servoing System
Objective: To develop the first known deep learning-based photoacoustic visual servoing system utilizing point source localization and hybrid position-force control to track catheter tips in three dimensions in real-time. Methods: We integrated either object detection or instance segmentation-based localization with hybrid position-force control to create our novel system. Cardiac catheter tips were then tracked across distances of 40 mm in a plastisol phantom and 25-64 mm in an in vivo swine in real-time in nine visual servoing trials total. Results: Object detection-based localization identified the cardiac catheter tip in 88.0-91.7% and 66.7-70.4% of phantom and in vivo channel data frames, respectively. Instance segmentation detection rates ranged 86.4-100.0% in vivo. These catheter tips were tracked with errors as low as 0.5 mm in phantom trials and 0.8 mm in the in vivo trials. The mean inference times were ≥145.3 ms and ≥516.3 ms with object detection-based and instance segmentation-based point source localization, respectively. Hybrid position-force control system enabled contact with the imaging surface during ≥99.43% of each visual servoing trial. Conclusion: Our novel deep learning-based photoacoustic visual servoing system was successfully demonstrated. Object detectionbased localization operated with inference times that are more suitable for real-time implementations while instance segmentation had lower tracking errors. Significance: After implementing suggested optimization modifications, our novel system has the potential to track catheter tips, needle tips, and other surgical tool tips in real-time during surgical and interventional procedures.
- 22 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 72 (6) , 786-805
Deep Learning to Localize Photoacoustic Sources in Three Dimensions: Theory and Implementation
Surgical tool tip localization and tracking are essential components of surgical and interventional procedures. The cross sections of tool tips can be considered as acoustic point sources to achieve these tasks with deep learning applied to photoacoustic channel data. However, source localization was previously limited to the lateral and axial dimensions of an ultrasound transducer. In this paper, we developed a novel deep learning-based three-dimensional (3D) photoacoustic point source localization system using an object detection-based approach extended from our previous work. In addition, we derived theoretical relationships among point source locations, sound speeds, and waveform shapes in raw photoacoustic channel data frames. We then used this theory to develop a novel deep learning instance segmentation-based 3D point source localization system. When tested with 4,000 simulated, 993 phantom, and 1,983 ex vivo channel data frames, the two systems achieved F1 scores as high as 99.82%, 93.05%, and 98.20%, respectively, and Euclidean localization errors (mean ± one standard deviation) as low as 1.46±1.11 mm, 1.58±1.30 mm, and 1.55±0.86 mm, respectively. In addition, the instance segmentation-based system simultaneously estimated sound speeds with absolute errors (mean ± one standard deviation) of 19.22±26.26 m/s in simulated data and standard deviations ranging 14.6-32.3 m/s in experimental data. These results demonstrate the potential of the proposed photoacoustic imaging-based methods to localize and track tool tips in three dimensions during surgical and interventional procedures.
- 22 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 72 (6) , 786-805
Deep Learning to Localize Photoacoustic Sources in Three Dimensions: Theory and Implementation
Surgical tool tip localization and tracking are essential components of surgical and interventional procedures. The cross sections of tool tips can be considered as acoustic point sources to achieve these tasks with deep learning applied to photoacoustic channel data. However, source localization was previously limited to the lateral and axial dimensions of an ultrasound transducer. In this paper, we developed a novel deep learning-based three-dimensional (3D) photoacoustic point source localization system using an object detection-based approach extended from our previous work. In addition, we derived theoretical relationships among point source locations, sound speeds, and waveform shapes in raw photoacoustic channel data frames. We then used this theory to develop a novel deep learning instance segmentation-based 3D point source localization system. When tested with 4,000 simulated, 993 phantom, and 1,983 ex vivo channel data frames, the two systems achieved F1 scores as high as 99.82%, 93.05%, and 98.20%, respectively, and Euclidean localization errors (mean ± one standard deviation) as low as 1.46±1.11 mm, 1.58±1.30 mm, and 1.55±0.86 mm, respectively. In addition, the instance segmentation-based system simultaneously estimated sound speeds with absolute errors (mean ± one standard deviation) of 19.22±26.26 m/s in simulated data and standard deviations ranging 14.6-32.3 m/s in experimental data. These results demonstrate the potential of the proposed photoacoustic imaging-based methods to localize and track tool tips in three dimensions during surgical and interventional procedures.
- Open Access
- 14 November 2024
- Published by SPIE-Intl Soc Optical Eng in Biophotonics Discovery
- Vol. 2 (01) , 1-19
Impact of skin tone on target size detectability in photoacoustic breast imaging
SignificancePhotoacoustic imaging has the potential to improve non-invasive breast cancer diagnosis. However, illumination through the skin introduces a skin tone bias, as greater melanin content increases optical absorption and can create acoustic clutter, reducing the visibility of various target sizes.AimWe investigate the impact of skin tone bias as a function of target sizes in three photoacoustic image reconstruction methods: fast Fourier transform (FFT)-based reconstruction, delay-and-sum (DAS) beamforming, and short-lag spatial coherence (SLSC) beamforming.MethodsThe three imaging methods were applied to channel data from multidomain simulations with 757, 800, and 1064 nm wavelengths, 11 target sizes (ranging 0.5 to 3 mm), 18 skin tones [ranging individual typology angles (ITAs), of –54 deg to 60 deg], and a previously validated realistic 3D breast model. The signal-to-noise ratio (SNR) and generalized contrast-to-noise ratio (gCNR) were measured to assess visibility.ResultsWith 757 and 800 nm wavelengths, targets underlying dark skin tones (ITA range: −54 deg to −33 deg) with sizes ≤3 mm were poorly visible with ≤2.38 to 4.21 mean SNR and ≤0.46 to 0.74 mean gCNR, with smaller targets generally being more visible with lighter skin tones. A 1064-nm transmit wavelength improved visualization with DAS and SLSC beamforming, relative to both FFT reconstruction with a 1064 nm wavelength and DAS or SLSC beamforming with 757 and 800 nm wavelengths. When combined with SLSC beamforming, the 1064-nm wavelength offered the greatest improvements, enabling visualization of simulated target sizes ranging from 0.5 to 3 mm underlying very light (ITA = 60 deg) to dark (ITA=−54 deg) skin tones, with mean SNR ≤10.01 and mean gCNR ≤1. Visualization of simulated vessel structures derived from in vivo photoacoustic images was consistent with simulation-based target size expectations.ConclusionsResults are promising for advancing next-generation photoacoustic imaging systems for breast cancer diagnosis across the range of skin tones represented in healthcare systems throughout the world.
- 14 November 2024
- Published by SPIE-Intl Soc Optical Eng in Biophotonics Discovery
- Vol. 2 (01) , 1-19
Open AccessImpact of skin tone on target size detectability in photoacoustic breast imaging
SignificancePhotoacoustic imaging has the potential to improve non-invasive breast cancer diagnosis. However, illumination through the skin introduces a skin tone bias, as greater melanin content increases optical absorption and can create acoustic clutter, reducing the visibility of various target sizes.AimWe investigate the impact of skin tone bias as a function of target sizes in three photoacoustic image reconstruction methods: fast Fourier transform (FFT)-based reconstruction, delay-and-sum (DAS) beamforming, and short-lag spatial coherence (SLSC) beamforming.MethodsThe three imaging methods were applied to channel data from multidomain simulations with 757, 800, and 1064 nm wavelengths, 11 target sizes (ranging 0.5 to 3 mm), 18 skin tones [ranging individual typology angles (ITAs), of –54 deg to 60 deg], and a previously validated realistic 3D breast model. The signal-to-noise ratio (SNR) and generalized contrast-to-noise ratio (gCNR) were measured to assess visibility.ResultsWith 757 and 800 nm wavelengths, targets underlying dark skin tones (ITA range: −54 deg to −33 deg) with sizes ≤3 mm were poorly visible with ≤2.38 to 4.21 mean SNR and ≤0.46 to 0.74 mean gCNR, with smaller targets generally being more visible with lighter skin tones. A 1064-nm transmit wavelength improved visualization with DAS and SLSC beamforming, relative to both FFT reconstruction with a 1064 nm wavelength and DAS or SLSC beamforming with 757 and 800 nm wavelengths. When combined with SLSC beamforming, the 1064-nm wavelength offered the greatest improvements, enabling visualization of simulated target sizes ranging from 0.5 to 3 mm underlying very light (ITA = 60 deg) to dark (ITA=−54 deg) skin tones, with mean SNR ≤10.01 and mean gCNR ≤1. Visualization of simulated vessel structures derived from in vivo photoacoustic images was consistent with simulation-based target size expectations.ConclusionsResults are promising for advancing next-generation photoacoustic imaging systems for breast cancer diagnosis across the range of skin tones represented in healthcare systems throughout the world.
- Open Access
- 13 November 2023
- Published by SPIE-Intl Soc Optical Eng in Journal of Biomedical Optics
- Vol. 29 (S1) , S11505-21
Deep learning in vivo catheter tip locations for photoacoustic-guided cardiac interventions
Cardiac interventional procedures are often performed to diagnose and treat cardiac arrhythmias (e.g., 18,000 to 44,500 cardiac catheter ablation procedures have been performed annually in the United States1). These procedures generally require catheter delivery from an insertion point in the thigh to the heart via the femoral vein. One of the most serious and potentially life-threatening complications of catheter ablations is the risk of cardiac perforation,1 which can be minimized with state-of-the-art catheter tip visualization methods.- 13 November 2023
- Published by SPIE-Intl Soc Optical Eng in Journal of Biomedical Optics
- Vol. 29 (S1) , S11505-21
Open AccessDeep learning in vivo catheter tip locations for photoacoustic-guided cardiac interventions
Cardiac interventional procedures are often performed to diagnose and treat cardiac arrhythmias (e.g., 18,000 to 44,500 cardiac catheter ablation procedures have been performed annually in the United States1). These procedures generally require catheter delivery from an insertion point in the thigh to the heart via the femoral vein. One of the most serious and potentially life-threatening complications of catheter ablations is the risk of cardiac perforation,1 which can be minimized with state-of-the-art catheter tip visualization methods. - Open Access
- 21 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 69 (6) , 2098-2114
Theoretical Framework to Predict Generalized Contrast-to-Noise Ratios of Photoacoustic Images With Applications to Computer Vision
The successful integration of computer vision, robotic actuation, and photoacoustic imaging to find and follow targets of interest during surgical and interventional procedures requires accurate photoacoustic target detectability. This detectability has traditionally been assessed with image quality metrics, such as contrast, contrast-to-noise ratio, and signal-to-noise ratio (SNR). However, predicting target tracking performance expectations when using these traditional metrics is difficult due to unbounded values and sensitivity to image manipulation techniques like thresholding. The generalized contrast-to-noise ratio (gCNR) is a recently introduced alternative target detectability metric, with previous work dedicated to empirical demonstrations of applicability to photoacoustic images. In this article, we present theoretical approaches to model and predict the gCNR of photoacoustic images with an associated theoretical framework to analyze relationships between imaging system parameters and computer vision task performance. Our theoretical gCNR predictions are validated with histogram-based gCNR measurements from simulated, experimental phantom, ex vivo , and in vivo datasets. The mean absolute errors between predicted and measured gCNR values ranged from ${3.2} \times {10}^{-{3}}$ to ${2.3} \times {10}^{-{2}}$ for each dataset, with channel SNRs ranging −40 to 40 dB and laser energies ranging 0.07 $\mu \text{J}$ to 68 mJ. Relationships among gCNR, laser energy, target and background image parameters, target segmentation, and threshold levels were also investigated. Results provide a promising foundation to enable predictions of photoacoustic gCNR and visual servoing segmentation accuracy. The efficiency of precursory surgical and interventional tasks (e.g., energy selection for photoacoustic-guided surgeries) may also be improved with the proposed framework.
- 21 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 69 (6) , 2098-2114
Open AccessTheoretical Framework to Predict Generalized Contrast-to-Noise Ratios of Photoacoustic Images With Applications to Computer Vision
The successful integration of computer vision, robotic actuation, and photoacoustic imaging to find and follow targets of interest during surgical and interventional procedures requires accurate photoacoustic target detectability. This detectability has traditionally been assessed with image quality metrics, such as contrast, contrast-to-noise ratio, and signal-to-noise ratio (SNR). However, predicting target tracking performance expectations when using these traditional metrics is difficult due to unbounded values and sensitivity to image manipulation techniques like thresholding. The generalized contrast-to-noise ratio (gCNR) is a recently introduced alternative target detectability metric, with previous work dedicated to empirical demonstrations of applicability to photoacoustic images. In this article, we present theoretical approaches to model and predict the gCNR of photoacoustic images with an associated theoretical framework to analyze relationships between imaging system parameters and computer vision task performance. Our theoretical gCNR predictions are validated with histogram-based gCNR measurements from simulated, experimental phantom, ex vivo , and in vivo datasets. The mean absolute errors between predicted and measured gCNR values ranged from ${3.2} \times {10}^{-{3}}$ to ${2.3} \times {10}^{-{2}}$ for each dataset, with channel SNRs ranging −40 to 40 dB and laser energies ranging 0.07 $\mu \text{J}$ to 68 mJ. Relationships among gCNR, laser energy, target and background image parameters, target segmentation, and threshold levels were also investigated. Results provide a promising foundation to enable predictions of photoacoustic gCNR and visual servoing segmentation accuracy. The efficiency of precursory surgical and interventional tasks (e.g., energy selection for photoacoustic-guided surgeries) may also be improved with the proposed framework.
- 1 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 24-27
GPU implementation of coherence-based photoacoustic beamforming for autonomous visual servoing
Visual servoing is a promising technique for surgical tool tip tracking and automated visualization of photoacoustic targets during interventional procedures. However, visual servoing in photoacoustic imaging is challenged by the trade-off between laser safety limits and the energy needed for reliable segmentation. Short-lag spatial coherence (SLSC) imaging has the potential to overcome this challenge by improving the signal quality of images acquired with low laser energies. This study introduces the first known GPU-based real-time implementation of SLSC imaging for photoacoustic imaging and applies this real-time algorithm to enhance segmentations for visual servoing. Results with ex vivo bovine tissue demonstrate that SLSC imaging recovers signals obtained with low energy (i.e., ≤ 268 µJ) with a signal-to-noise ratio (SNR) of 11.2 ± 2.4, compared to a SNR of 3.5 ± 0.8 with conventional delay-and-sum (DAS) imaging. When energies were lower than the safety limit for skin (i.e., 400 µJ for 900 nm wavelength), real-time SLSC imaging produced lower errors during fiber tracking tasks (i.e., accuracies of 0.67 ± 0.42 mm and 4.91 ± 6.01 mm with SLSC and DAS, respectively). Similarly, for probe centering tests, the tracking error with SLSC and DAS imaging was 0.85 mm ±0.44 and 1.05 ± 0.30 mm, respectively. These results are promising for complicated visual servoing tasks in high-noise environments.
- 1 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 24-27
GPU implementation of coherence-based photoacoustic beamforming for autonomous visual servoing
Visual servoing is a promising technique for surgical tool tip tracking and automated visualization of photoacoustic targets during interventional procedures. However, visual servoing in photoacoustic imaging is challenged by the trade-off between laser safety limits and the energy needed for reliable segmentation. Short-lag spatial coherence (SLSC) imaging has the potential to overcome this challenge by improving the signal quality of images acquired with low laser energies. This study introduces the first known GPU-based real-time implementation of SLSC imaging for photoacoustic imaging and applies this real-time algorithm to enhance segmentations for visual servoing. Results with ex vivo bovine tissue demonstrate that SLSC imaging recovers signals obtained with low energy (i.e., ≤ 268 µJ) with a signal-to-noise ratio (SNR) of 11.2 ± 2.4, compared to a SNR of 3.5 ± 0.8 with conventional delay-and-sum (DAS) imaging. When energies were lower than the safety limit for skin (i.e., 400 µJ for 900 nm wavelength), real-time SLSC imaging produced lower errors during fiber tracking tasks (i.e., accuracies of 0.67 ± 0.42 mm and 4.91 ± 6.01 mm with SLSC and DAS, respectively). Similarly, for probe centering tests, the tracking error with SLSC and DAS imaging was 0.85 mm ±0.44 and 1.05 ± 0.30 mm, respectively. These results are promising for complicated visual servoing tasks in high-noise environments.
- Open Access
- 5 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 39 (4) , 1015-1029
In Vivo Demonstration of Photoacoustic Image Guidance and Robotic Visual Servoing for Cardiac Catheter-Based Interventions
Cardiac interventional procedures are often performed under fluoroscopic guidance, exposing both the patient and operators to ionizing radiation. To reduce this risk of radiation exposure, we are exploring the use of photoacoustic imaging paired with robotic visual servoing for cardiac catheter visualization and surgical guidance. A cardiac catheterization procedure was performed on two in vivo swine after inserting an optical fiber into the cardiac catheter to produce photoacoustic signals from the tip of the fiber-catheter pair. A combination of photoacoustic imaging and robotic visual servoing was employed to visualize and maintain constant sight of the catheter tip in order to guide the catheter through the femoral or jugular vein, toward the heart. Fluoroscopy provided initial ground truth estimates for 1D validation of the catheter tip positions, and these estimates were refined using a 3D electromagnetic-based cardiac mapping system as the ground truth. The 1D and 3D root mean square errors ranged 0.25-2.28 mm and 1.24-1.54 mm, respectively. The catheter tip was additionally visualized at three locations within the heart: (1) inside the right atrium, (2) in contact with the right ventricular outflow tract, and (3) inside the right ventricle. Lasered regions of cardiac tissue were resected for histopathological analysis, which revealed no laserrelated tissue damage, despite the use of 2.98 mJ per pulse at the fiber tip (379.2 mJ/cm2 fluence). In addition, there was a 19 dB difference in photoacoustic signal contrast when visualizing the catheter tip pre-and post-endocardial tissue contact, which is promising for contact confirmation during cardiac interventional procedures (e.g., cardiac radiofrequency ablation). These results are additionally promising for the use of photoacoustic imaging to guide cardiac interventions by providing depth information and enhanced visualization of catheter tip locations within blood vessels and within the beating heart.
- 5 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 39 (4) , 1015-1029
Open AccessIn Vivo Demonstration of Photoacoustic Image Guidance and Robotic Visual Servoing for Cardiac Catheter-Based Interventions
Cardiac interventional procedures are often performed under fluoroscopic guidance, exposing both the patient and operators to ionizing radiation. To reduce this risk of radiation exposure, we are exploring the use of photoacoustic imaging paired with robotic visual servoing for cardiac catheter visualization and surgical guidance. A cardiac catheterization procedure was performed on two in vivo swine after inserting an optical fiber into the cardiac catheter to produce photoacoustic signals from the tip of the fiber-catheter pair. A combination of photoacoustic imaging and robotic visual servoing was employed to visualize and maintain constant sight of the catheter tip in order to guide the catheter through the femoral or jugular vein, toward the heart. Fluoroscopy provided initial ground truth estimates for 1D validation of the catheter tip positions, and these estimates were refined using a 3D electromagnetic-based cardiac mapping system as the ground truth. The 1D and 3D root mean square errors ranged 0.25-2.28 mm and 1.24-1.54 mm, respectively. The catheter tip was additionally visualized at three locations within the heart: (1) inside the right atrium, (2) in contact with the right ventricular outflow tract, and (3) inside the right ventricle. Lasered regions of cardiac tissue were resected for histopathological analysis, which revealed no laserrelated tissue damage, despite the use of 2.98 mJ per pulse at the fiber tip (379.2 mJ/cm2 fluence). In addition, there was a 19 dB difference in photoacoustic signal contrast when visualizing the catheter tip pre-and post-endocardial tissue contact, which is promising for contact confirmation during cardiac interventional procedures (e.g., cardiac radiofrequency ablation). These results are additionally promising for the use of photoacoustic imaging to guide cardiac interventions by providing depth information and enhanced visualization of catheter tip locations within blood vessels and within the beating heart.