Multiscale Vascular Enhancement Filter Applied toIn VivoMorphologic and Functional Photoacoustic Imaging of Rat Ocular Vasculature

Abstract

Optical-resolution photoacoustic microscopy (OR-PAM) is used for in vivo imaging of a variety of albino and pigmented eyes taking advantages of requiring no exogenous dye, performing high-resolution imaging, and achieving morphologic and functional imaging at the same time. However, to accurately diagnose the ophthalmic disease in the OR-PAM images, vascular enhancement algorithms are necessary for extracting vessels and quantifying them correctly. Vascular enhancement algorithms developed for other imaging technologies, are not suitable to be used for OR-PAM, because of the underlying differences in the physics of the formation of images. In this study, a new vascular enhancement algorithm called photoacoustic imaging vasculature enhancement filter (PAIVEF) is proposed, which not only enhances vasculature including micro-vessels signals, suppresses noise signals effectively, but also achieves highly sensitive and accurate enhancement of the vasculature within a large depth range in and out of the system's depth of focus (DOF). Using the PAIVEF, the morphologic and functional 3D images of the whole rat's ocular anterior vasculature segment was displayed simultaneously for a depth range of ∼0.6 mm, which was ∼7 times of the system's DOF. This study paves the way for the application of OR-PAM technology in ophthalmic disease research.

Versatile Single-Element Ultrasound Imaging Platform using a Water-Proofed MEMS Scanner for Animals and Humans

“…Third, global 3D volumetric imaging cannot be performed seamlessly due to motion artifacts between consecutive B-mode images (e.g., respiration and heartbeat). These problems can be overcome by image processing methods, such as compressive sensing, and electrocardiogram (ECG) gating technology [35][36][37][38] . In addition, the continuously steered mirror also could lead the motion artifact because of the directivity of acoustic.…”

Deep Learning Enables Superior Photoacoustic Imaging at Ultralow Laser Dosages

“…The loss function was minimized, and the CNN‐related parameters were continuously updated to obtain the best training models. Corresponding to Outputs 1–3, Ground truths 1–3 were full sampling images obtained at the ANSI limit per pulse laser energy at 532 nm, full sampling images obtained at ANSI limit per pulse laser energy at 560 nm, and Ground truth 1 filtered by the PAIVEF method, [ 26 ] respectively. Notably, Ground truth 3 was Ground truth 1 processed by the PAIVEF method to obtain enhanced image quality with a conventional non‐deep‐learning strategy.…”

“…The PAIVEF method is an image processing algorithm based on Frangi's filter and is verified as an optimal vasculature enhancement filter which significantly improves image quality and causes less distortion in photoacoustic imaging. [ 26 ] Similar to training, during testing, Input 1 and Input 2 were obtained by sectioning the original images (532 and 560 nm) into small subsections, which were then used as inputs into the MT‐RDN to obtain Output 1, Output 2, and Output 3. The subsections in each output were stitched together to obtain Recon 1, Recon 2, and Recon 3.…”

“…Various image and signal processing algorithms have been developed in OR‐PAM to improve the image quality, such as image denoising, [ 22,23 ] compressed sensing, [ 24 ] and vascular filtering. [ 25,26 ] While these algorithms have progressively improved the quality of OR‐PAM imaging successfully, the extent of the improvement using image‐based postprocessing methods has been limited. The image denoising algorithms [ 23 ] may work well in simulations or phantoms, but improving the imaging quality of living samples is extremely difficult due to the complexity of biological tissues.…”

Assessment of hessian-based Frangi vesselness filter in optoacoustic imaging

“…In OA mesoscopy and microscopy, the HFV filter has been widely employed for vasculature enhancement [ [20] , [21] , [22] , 5 , 23 ] and quantification [ 24 , 25 ]. Furthermore, hybrid enhancement techniques, such as the combination of a Gabor filter and either a standard [ 26 ] or modified Frangi vesselness filter [ 9 , 27 ] have been also employed to improve background noise suppression and small vessel contrast. In OA macroscopy, the HFV filter has been applied in clinical and pre-clinical studies on either fully reconstructed images [ 3 , 7 , 8 ] or separately on both positive and negative components of the reconstructed image [ 2 ].…”