Accelerating Whole-Sample Polarization-Resolved Second Harmonic Generation imaging in Mammary Gland Tissue via Generative Adversarial Networks

<p>Authors:</p> <p>Arash Aghigh, Jysiane Cardot, Melika Saadat Mohammadi, Ga&euml;tan Jargot, Heide Ibrahim, Isabelle Plante, Fran&ccedil;ois L&eacute;gar&eacute;</p> <p>Affiliations:</p> <p>&nbsp; &nbsp; 1. &nbsp; &nbsp;Centre &Eacute;nergie Mat&eacute;riaux T&eacute;l&eacute;communications, Institut National de la Recherche Scientifique, Varennes, Qu&eacute;bec, Canada.<br>&nbsp; &nbsp; 2. &nbsp; &nbsp;Centre Armand-Frappier Sant&eacute; Biotechnologie, Institut National de la Recherche Scientifique, Laval, Qu&eacute;bec, Canada.</p> <p>Corresponding Author:</p> <p>Arash Aghigh, arash.aghigh@inrs.ca</p> <p>Description:</p> <p>This dataset accompanies the research on improving whole-sample Polarization-Resolved Second Harmonic Generation (P-SHG) imaging in mammary gland tissue using Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN). The novel approach significantly reduces imaging time while maintaining high image quality and analytical accuracy, demonstrating a reduction in imaging time by more than 95%. This method also minimizes laser-induced photodamage, lowers costs of optical components, and increases the accessibility and applicability of P-SHG imaging in various fields.</p> <p>Keywords:</p> <p>Polarization-Resolved Second Harmonic Generation, P-SHG, Generative Adversarial Networks, GAN, ESRGAN, Mammary Gland Imaging, Super-Resolution, Image Upscaling, Deep Learning, Biomedical Imaging</p> <p>Funding Information:</p> <p>&nbsp; &nbsp; &bull; &nbsp; &nbsp;Canada Foundation for Innovation<br>&nbsp; &nbsp; &bull; &nbsp; &nbsp;Fonds de recherche du Qu&eacute;bec&ndash;Nature et technologies<br>&nbsp; &nbsp; &bull; &nbsp; &nbsp;Natural Sciences and Engineering Research Council of Canada<br>&nbsp; &nbsp; &bull; &nbsp; &nbsp;New Frontiers Research Fund<br>&nbsp; &nbsp; &bull; &nbsp; &nbsp;NSERC CREATE program (scholarship for Arash Aghigh)</p> <p>Related Identifiers:</p> <p>&nbsp;&nbsp;&nbsp; &bull; &nbsp; &nbsp;GitHub repository for ChaiNNer program: https://github.com/chaiNNer-org/chaiNNer<br>&nbsp; &nbsp; &bull; &nbsp; &nbsp;Download links for models used: https://openmodeldb.info</p> <p>Additional Information:</p> <p>Animal studies were conducted according to the procedures provided by the Canadian Council on Animal Care. The protocol (2005-02) was reviewed and approved by the Institutional Committee for Animal Protection of the Laboratoire National de Biologie Exp&eacute;rimentale (LNBE), the animal facilities based at the Institut National de Recherche Scientifique (INRS).</p>