CycleGAN network for photoacoustic histology

<p>An unsupervised deep learning algorithm based on cycle-consistent generative adversarial networks (CycleGAN&nbsp;) converted the UV-PAM images into H&amp;E-like pseudocolor histologic images, allowing the pathologists to readily identify the cancerous features following existing pattern-recognition parameters.&nbsp;Unlike supervised deep learning methods such as generational adversarial networks (GAN), the&nbsp;unsupervised deep learning method based on CycleGAN&nbsp;does not require coupled pairs of stained and unstained images. It avoids the need for well-aligned UV-PAM and H&amp;E-stained images for neural network training, which can be challenging to acquire due to artifacts caused by sample preparation-induced morphology changes.</p>