Deconvolution-based approach for super-resolution photoacoustic imaging

Abstract

Recently, the microbubble-localization technique was able to increase spatial resolution by multifold in ultrasound (US) imaging, overcoming the acoustic diffraction limit. In this study, we applied the super-resolution approach to the photoacoustic (PA) imaging with a mid-frequency medical imaging array transducer. One of the biggest challenges in this application is relatively low PA response from individual contrast agent particles to the limited laser fluence under safety guidelines. The increased concentration of contrast agent enhances PA signal, but it results in convolved and blurred PA signals due to particle clumping. The conventional localization technique likely discards such blurred PA signals and thus requires a large number of imaging frames. Alternatively, a deconvolution method can be utilized to directly identify individual agent particles from the blurred PA signals. In-vitro experiments were performed using the PA sequence programmed US platform with a linear array transducer centered at 5 MHz. A pulsed laser of 5 ns long at 10 Hz tuned at 800 nm was synchronized to the US system. Totally 8,000 PA frames were acquired from micron-sized metal beads flowing in a polyethylene tube. Richardson-Lucy deconvolution method successfully identified individual particles from the blurred PA signals generated from clumped beads at high concentration.