Context:
3D fast ultrasound imaging has emerged in the last year researches but still suffers from its poor image quality. Indeed, using plane or diverging waves does not permit to insonify the medium with sufficient energy at each point to get a good signal-to-noise ratio (SNR), contrast-to-noise-ratio (CNR) or even resolutions. On the other hand, coded excitation is currently used to increase signal-to-noise ratio and penetration depth. The final objective is to combine 3D fast/ultrafast imaging with coded excitation to achieve better image quality at a high acquisition rate.
Approaches:
The axial resolution of an ultrasound imaging system is inversely proportional to the bandwidth of the transmitted signal. When using conventional pulsing, the impulse response of the transducer and the excitation signal together determine the shape of the emitted pulse and its bandwidth.
A first technique to increase the ultrasound image resolution is to use the "Thomson's multitaper" approach. With the aim of improving resolution and contrast while reducing speckle noise and preserving frame rate, we proposed to combine Thomson’s multitaper method and coherent plane-wave compounding technique to produce technique that takes advantage of coherent and incoherent approaches [Toulemonde-2015-Ultrasonics][Basset-2015-IRBM].
The second way is to virtually increase the transducer frequency band. Resolution enhancement compression (REC) is a coding technique that increases signal energy in transition frequency bands, where the energy transduction of the ultrasound probe is less efficient. We combined REC with coherent plane-wave compounding to improve image quality in high-frame-rate modes. A linear, amplitude modulated, chirp is transmitted inside a medium and the received echoes must be compressed to restore a good resolution in the final image. This pulse compression is achieved by a matched filter which is a linear time-invariant filter that maximizes the echo Signal-to-Noise-Ratio (eSNR) in presence of white Gaussian noise. Experimental results, on ex vivo rabbit livers, showed that an effective implementation of REC on a research scanner using CPWC is possible [Benane-2018-IEEE TUFFC].
This technique has been extended from 2D to 3D using the Four Verasonics VantageTM 256 systems (Verasonics, Redmond, Washington, United States). They allow to transmit and acquire the 1024 channels. Each system controls 8-by-32 elements of a total of 32-by-32. The experimental results shown hereafter were conducted using a Gammex phantom model 410 SCG (Gammex, Middleton, Wisconsin, United States). An acquisition scheme with a pulse repetition frequency of 100 Hz was used based on 7 angle transmissions on each axis, resulting in 49 compound images. Two different transmission signals are compared: (i) a standard 2.5 sinusoidal cycles, and (ii) a 6.7 μs chirp transmission from 2.1 to 3.9 MHz
Figure 1: 3D Resolution enhancement compression (3D REC with chirp signal) on a Gammex phantom
Another possibility to increase the resolution, and at the same time to increase the contrast-to-tissue ratio (CTR) in contrast imaging or the signal-to-noise ratio (SNR) in tissue harmonic imaging, is to use multi-pulse transmission techniques [Lin-2015-Ultrasonic Imaging]. The proposed design can be used to predict the relative amplitude of the nonlinear components in each frequency band and to design new transmission sequences to increase or decrease some nonlinear components.
Finally, a third approach was proposed for coded emission imaging. The technique is based on the simultaneous emission of time-encoded plane waves (SCE on the Figure 2). An advanced direct linear model that links the received backscattered echoes to the Tissue Reflectivity Function (TRF) was proposed. To estimate TRF, a regularized inverse problem was solved using the prior information of the TRF sparsity [Bujoreanu-2017-IEEE ICASSP].
Figure 2: Principle of the Simultaneous Coded Emission (SCE) vs Coherent Compounding (CC)
Figure 3: Results obtained on a hypoechoic cyst using Simultaneous Coded Emission (System UlaOp 64) [Bujoreanu-2017-IEEE IUS].