P J Cao

University of Southern California, Los Ángeles, California, United States

Are you P J Cao?

Claim your profile

Publications (6)3.58 Total impact

  • C.-H. Hu · K.A. Snook · X.-C. Xu · J.T. Yen · K.K. Shung · P.-J. Cao
    [Show abstract] [Hide abstract]
    ABSTRACT: We have developed two digital beamformers for eye imaging using 30-50 MHz transducer arrays: a 6-channel digital beamformer for a six-ring annular array ultrasound transducer (43MHz) and a 16-channel digital beamformer for a linear array ultrasound transducer (30MHz) of 48-64 elements. The hardware architecture of both designs allows the flexibility for multiple transmit focusing, dynamic receive focusing, amplitude apodization, and dynamic aperture control. The beamformer uses Xilinx Virtex-E FPGA chips (XCV300E, Xilinx, Inc.). Individual channel delays are applied using a combination of coarse and fine delays. The coarse delays are integer multiples of the clock periods achieved by using a first in first out (FIFO) structure, and the fine delays are applied with interpolation of a 4-tap Fractional Delay (FD) FIR filter. The dynamic receive focusing, with the smallest increments of 0.5mm, is implemented based on a 16-state state machine which updates delay values and the FD filter coefficients according to the depth of the echoes. The aperture apodization in the digital beam former for linear array is set according to the focusing zone. For near zone, middle zone, far zone focusing, aperture sizes are set to 8,12,and 16 elements respectively. Cross-sectional wire target images using a 20 μm diameter Tungsten wires from both annular and linear array are shown. The images of an excised pig eye clearly show the fine structures of the anterior segment of eye including cornea and lens.
    No preview · Article · Jan 2004
  • [Show abstract] [Hide abstract]
    ABSTRACT: Blood flow disturbance by a severe eccentric stenosis (∼ 70 %) was investigated in a polystyrene tube (9.5 mm of inner diameter) of a mock flow loop. A modified Aloka 280 system with a 7.5 MHz linear array transducer was used to collect the RF (radio frequency) and video signals of B-mode images. Pulsatile flow was generated and controlled to observe the 'black hole (BH)' phenomenon, a hypo-echoic zone at the center stream, upstream of the stenosis. The BH from porcine whole blood was disturbed on the stenotic site but reappeared further downstream at 20 beats/min (bpm). At higher stroke rates (40 and 60 bpm), the BH was not observed downstream but a hyper-echoic parabolic profile appeared at each systole showing the periodic spatial variation further downstream up to 10 diameters. These bright profiles seemed to be due to enhancement of rouleaux formation by flow acceleration. The ultrasonic power from RBC suspensions under steady flow showed a strong echoic horizontal line downstream of a stenosis and its position and strength varied with flow speed. It seemed to be related with the boundary between the two disturbed flow zones. In conclusion, a severe eccentric stenosis disturbed the blood flow and red cell structures, so that the backscatter from porcine whole blood varied with flow speed, stroke rate, and position along the flow direction. Ultrasonic B-mode imaging may be a useful tool to visualize the blood flow and to study hemodynamics.
    No preview · Article · Jan 2004 · Key Engineering Materials
  • [Show abstract] [Hide abstract]
    ABSTRACT: A 30 MHz annular array imaging system and beamforming algorithm were developed as an intermediate step to improve focusing, and therefore image resolution, throughout the image field via dynamic focusing in 2D. Transmit and receive beamforming was accomplished with the full aperture by applying delays to each element. The annular array with six equal area elements was designed and fabricated for use with the beamformer. The array incorporated a fine-grain lead titanate, which provided good sensitivity with reduced lateral coupling. Each transducer element provided over 40% bandwidth and less than 18 dB insertion loss. The hardware system included six channels, each incorporating a pulser/receiver, time-gain compensation (TGC), and an amplifier. In this study, during reception, the RF data for each channel was digitized sequentially, due to the limited number of 500 MHz A/D converters, stored for off-line calculation and displayed with software. A model was used to evaluate the point-spread function of the system throughout the image field. Simulation showed that the point spread function in the axial direction was about 65 μm (-6 dB) and 105 μm (20 dB) throughout a distance of 10 mm to 18 mm. The lateral resolution was 150 μm (-6 dB) from 10-18 mm, whereas the -20 dB lateral resolution increased from 150 μm to 300 μm in the same range. Tests with an 8 μm diameter tungsten wire phantom showed an axial resolution of 55 μm (-6 dB) and 145 μm (-20 dB) at a depth of 10 mm to 17 mm. The -6 dB lateral resolution increased from 100 μm to 240 μm in the same range. The -20 dB lateral resolution was approximately 440 μm to 640 μm from 10 mm to 17 mm. The simulated and experimental results showed that improved resolution can be attained using cross-fade method. Preliminary images of an excised eye were acquired with the imaging system in vitro.
    No preview · Conference Paper · Nov 2002
  • D G Paeng · P J Cao · K.Kirk Shung
    [Show abstract] [Hide abstract]
    ABSTRACT: Although a number of recent studies have demonstrated that the echogenicity of blood varies as a function of time under pulsatile flow, the fundamental mechanisms responsible for it are still uncertain. To better understand this phenomenon, the Doppler power from porcine blood and polystyrene microsphere suspensions was measured at the center of the tube as functions of two crucial parameters, flow velocity and stroke rate (for pulsatile flow), under steady and pulsatile flow in a mock flow loop. In the present study, the experimental results were obtained with a 10-MHz pulsed Doppler system with a frequency response estimated more accurately by electronic injection, and validated by comparing to the radiofrequency (RF) signal acquired from the same Doppler instrument. The results show that the Doppler power from microspheres and porcine red blood cell (RBC) suspensions did not vary appreciably (< 2 dB), with either the speed or stroke rate (for pulsatile flow only) under steady and pulsatile flow. It was found that the Doppler power from porcine whole blood under steady flow decreased with the speed by approximately 13 dB from 3 to 33 cm/s and was only 3 dB higher than that from RBC suspension at 33 cm/s, suggesting minimal RBC aggregation in whole blood at this speed. The apparent cyclic variation from whole blood was observed at 20 and 40 beats/min (BPM). The cyclic variation became more obvious as the speed and stroke rate decreased. The mean Doppler power over a cycle increased as the peak speed decreased. The Doppler power reached a maximum near peak systole and a minimum at late diastole at the center of the tube. This pattern cannot be explained by RBC aggregation due to the shear rate alone, and may be attributed to acceleration and deceleration along with aggregation. The cyclic variation was not observed at 60 BPM, probably because of a lack of time for aggregation to occur.
    No preview · Article · Sep 2001 · Ultrasound in Medicine & Biology
  • W.H. Chen · P.J. Cao · E. Maione · T. Ritter · K.K. Shung
    [Show abstract] [Hide abstract]
    ABSTRACT: The effect of cables that connect the transducer and an imaging system can no longer be ignored when the ultrasound frequency is higher than 20MHz. The length of transmission line between transducer and electronic system critically affects the amplitude and bandwidth of pulse transmission. We report an evaluation method for assessing the effects of the cables and results obtained from this method. A time domain PSpice simulation model including high frequency impulse generator and receiver, a transducer simulated with a Mason model and a lossy transmission line has been developed. The model allows the simulation of the echo pulse shape at the input of the receiver. The investigation is carried out with a transducer of center frequency 45MHz and 48% bandwidth. The effect of cable length on the amplitude, space pulse length and bandwidth of received pulse is studied. Moreover, the effect of impedance tuning is also studied between the system and transducer when the cable is not used. Both experimental and simulation results show that the system performance is optimized by a cable length of 60cm to achieve higher sensitivity. If a suitable impedance matching method is applied between the system and the transducer without the cable connected, the system performance in sensitivity would be better than that with a cable connected. However, the results also show that achieving a highest sensitivity would increase pulse length. The results show that optimizing high frequency pulse transmission with high sensitivity and acceptable pulse shape, the pulser and receiver should be placed as close as possible to transducer with suitable impedance matching. The results also suggest the shorted pulse length could occur at a certain cable length but the result is also dependent on the original pulse shape
    No preview · Conference Paper · Feb 2001
  • P J Cao · D G Paeng · K K Shung
    [Show abstract] [Hide abstract]
    ABSTRACT: The "black hole" phenomenon was further investigated with porcine whole blood under pulsatile flow conditions in a straight rigid tube 120 cm long and of 0.95 cm diameter. A modified Aloka 280 commercial scanner with a 7.5 MHz linear array was used to collect the radio frequency (RF) signal of backscattering echoes from the blood inside the tube. The transducer was located downstream from the entrance and parallel to the longitudinal direction of the tube. The experimental results showed that higher hematocrits enhanced the black hole phenomenon, leading to a more apparent and larger diameter black hole. The black hole was not apparent at hematocrits below 23%. The highest hematocrit used in the experiment was 60%. Beat rates of 20, 40 and 60 beats per minute (bpm) were used, and the black hole became weaker in amplitude and smaller in diameter when the peak flow velocity was increased at each beat rate. These results are consistent with the suggestion in previous work that the black hole arises from insufficient aggregation of red blood cells (RBCs) at the center of the tube because of the low shear rate. At 20 and 40 bpm, the peak flow velocity ranges were 10 approximately 25 cm/s and 18 approximately 27 cm/s, respectively. The black hole was very clear at the minimal peak flow velocity but almost disappeared at the maximal velocities for each beat rate. At 60 bpm, experiments were only performed at one peak flow velocity of 31 cm/s and the black hole was clear. The results showed that the black hole was more pronounced at higher beat rates when the peak velocity was the same. This phenomenon cannot be explained by previous hypotheses. Acceleration seems to be the only flow parameter that varies at different beat rates when peak velocities are the same. Therefore, the influence of acceleration on the structural organization and orientation of RBC rouleaux might be another factor involved in the formation of the black hole in addition to the shear rate. As the entrance length was changed from 110 to 15 diameters (D) in seven steps at the hematocrit of 60%, it was found that a position farther downstream yielded a black hole with a greater contrast relative to the surrounding region, while the backscattering power at the central hypoechoic zone did not increase with increasing entrance length.
    No preview · Article · Feb 2001 · Biorheology

Publication Stats

67 Citations
3.58 Total Impact Points


  • 2002
    • University of Southern California
      • Department of Biomedical Engineering
      Los Ángeles, California, United States
  • 2001
    • Pennsylvania State University
      • Department of Bioengineering
      University Park, MD, United States