<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p><span class="break-words "><span><span dir="ltr">Dear all,</span></span></span></p>
    <p><span class="break-words "><span><span dir="ltr">We are pleased
            to announce the release of three richly annotated datasets
            for surgical video analysis: <br>
          </span></span></span></p>
    <p><span class="break-words "><span><span dir="ltr">1. Endoscapes is
            a dataset of 201 laparoscopic videos richly annotated for
            scene segmentation, object detection, and critical view of
            safety assessment. <br>
               arXiv: <a class="moz-txt-link-freetext" href="https://arxiv.org/abs/2312.12429">https://arxiv.org/abs/2312.12429</a><br>
               code: <a class="moz-txt-link-freetext" href="https://github.com/CAMMA-public/Endoscapes">https://github.com/CAMMA-public/Endoscapes</a></span></span></span></p>
    <p><span class="break-words "><span><span dir="ltr">2. SSG-VQA is a
            Visual Question Answering dataset on laparoscopic videos
            providing diverse, geometrically grounded, unbiased, and
            surgical action-oriented queries generated using surgical
            scene graphs.<br>
               arXiv: <a class="moz-txt-link-freetext" href="https://arxiv.org/abs/2312.10251">https://arxiv.org/abs/2312.10251</a> <br>
               </span></span></span><span class="break-words "><span><span
            dir="ltr"><span class="break-words "><span><span dir="ltr">code</span></span></span>:
            <a class="moz-txt-link-freetext" href="https://github.com/CAMMA-public/SSG-VQA">https://github.com/CAMMA-public/SSG-VQA</a></span></span></span></p>
    <p><span class="break-words "><span><span dir="ltr">3.
            Multibypass140 is a multicentric dataset of endoscopic video
            of laparoscopic Roux-en-Y gastric bypass surgery for
            research on multi-level surgical activity recognition,
            specifically phases, and steps.<br>
               arXiv: <a class="moz-txt-link-freetext" href="https://arxiv.org/abs/2312.11250">https://arxiv.org/abs/2312.11250</a><br>
               </span></span></span><span class="break-words "><span><span
            dir="ltr"><span class="break-words "><span><span dir="ltr">code</span></span></span>:
            <a class="moz-txt-link-freetext" href="https://github.com/CAMMA-public/MultiBypass140">https://github.com/CAMMA-public/MultiBypass140</a><br>
          </span></span></span></p>
    <p><span class="break-words "><span><span dir="ltr"> <br>
            Kind regards,<br>
          </span></span></span></p>
    <p><span class="break-words "><span><span dir="ltr">Nicolas Padoy</span></span></span></p>
    <pre class="moz-signature" cols="72">-- 
Nicolas Padoy
Professor of Computer Science, University of Strasbourg
Scientific Director, IHU Strasbourg
Director of CS & AI Research, IHU Strasbourg
Head of Research Group CAMMA 
Web: <a class="moz-txt-link-freetext" href="http://camma.u-strasbg.fr">http://camma.u-strasbg.fr</a></pre>
  </body>
</html>