Abstract
Photoacoustic tomography (PAT) is a non-ionizing imaging modality capable of acquiring high contrast and resolution images of optical absorption at depths greater than traditional optical imaging techniques. Practical considerations with instrumentation and geometry limit the number of available acoustic sensors and their “view” of the imaging target, which result in image reconstruction artifacts degrading image quality. Iterative reconstruction methods can be used to reduce artifacts but are computationally expensive. In this work, we propose a novel deep learning approach termed pixel-wise deep learning (Pixel-DL) that first employs pixel-wise interpolation governed by the physics of photoacoustic wave propagation and then uses a convolution neural network to reconstruct an image. Simulated photoacoustic data from synthetic, mouse-brain, lung, and fundus vasculature phantoms were used for training and testing. Results demonstrated that Pixel-DL achieved comparable or better performance to iterative methods and consistently outperformed other CNN-based approaches for correcting artifacts. Pixel-DL is a computationally efficient approach that enables for real-time PAT rendering and improved image reconstruction quality for limited-view and sparse PAT.
Highlights
Neuroimaging in small animals have played an essential role in preclinical research to provide physiological, pathological, and functional insights that are key for understanding and treating neurological diseases
We propose a novel deep learning approach termed Pixel-DL for limited-view and sparse Photoacoustic tomography (PAT) image reconstruction
We performed in silico experiments using training and testing data derived from multiple vasculature phantoms to compare Pixel-DL with conventional PAT image reconstruction methods and direct learned approaches (Post-DL and mDirect-DL)
Summary
Neuroimaging in small animals have played an essential role in preclinical research to provide physiological, pathological, and functional insights that are key for understanding and treating neurological diseases. Building an imaging system with these specifications is often prohibitively expensive, and in many in vivo applications such as neuroimaging, the sensor array typically can only www.nature.com/scientificreports partially enclose the tissue[29,30] These practical limitations result in sparse spatial sampling and limited-view of the photoacoustic waves emanating from the medium. Iterative methods are commonly employed to remove artifacts and improve image quality These methods use an explicit model of photoacoustic wave propagation and seek to minimize a penalty function that incorporates prior information[32,33,34]. Deep learning has the potential to be an effective and computationally efficient alternative to state-of-the-art iterative methods Having such a method would enable improved image quality, real-time PAT image rendering, and more accurate image interpretation and quantification
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
