Compressed sensing photoacoustic imaging based on discrete cosine model

To address the issues of high computational complexity and long reconstruction time caused by high sampling rates in photoacoustic imaging, this paper proposes a compressed sensing photoacoustic imaging method based on the discrete cosine model. This method adopts a sparse signal representation approach that does not require the construction of a sparse transform basis, using the discrete cosine domain photoacoustic forward model as the sensing matrix. By establishing the mapping relationship of the sensing matrix, the method links the sparse representation of photoacoustic signals with the distribution of photoacoustic sources in the discrete cosine domain, even with fewer channels and lower sampling rates. Furthermore, this paper discusses the optimization effect of an inverse compressed sensing-based photoacoustic image enhancement method on sparse channel photoacoustic imaging. Experimental results demonstrate that compared to traditional compressed sensing photoacoustic imaging methods that require the construction of a sparse transform basis, the proposed method not only achieves lower relative residuals but also significantly reduces computation time. Even with half the number of ultrasound transducers, high-quality and low-artifact photoacoustic images can still be obtained.