Skip to content

Commit

Permalink
Update index.html w/ TPAMI 2023 info
Browse files Browse the repository at this point in the history
  • Loading branch information
@jflalonde
jflalonde authored Jul 29, 2023
1 parent 9b39378 commit 6bfd54f
Showing 1 changed file with 18 additions and 19 deletions.
37 changes: 18 additions & 19 deletions deepcalib/index.html
View file
Original file line number Diff line number Diff line change
Expand Up @@ -50,8 +50,8 @@
<div class="col-md-12">
<div class="intro-message">
<h1></h1>
<h1>A Perceptual Measure for Deep Single Image <br/> Camera and Lens Calibration</h1>
<p>Image editing and compositing have become ubiquitous in entertainment, from digital art to AR and VR experiences. To produce beautiful composites, the camera needs to be geometrically calibrated, which can be tedious and requires a physical calibration target. In place of the traditional multi-images calibration process, we propose to infer the camera calibration parameters such as pitch, roll, field of view, and lens distortion directly from a single image using a deep convolutional neural network. We train this network using automatically generated samples from a large-scale panorama dataset, yielding competitive accuracy in terms of standard l2 error. However, we argue that minimizing such standard error metrics might not be optimal for many applications. In this work, we investigate human sensitivity to inaccuracies in geometric camera calibration. To this end, we conduct a large-scale human perception study where we ask participants to judge the realism of 3D objects composited with correct and biased camera calibration parameters. Based on this study, we develop a new perceptual measure for camera calibration and demonstrate that our deep calibration network outperforms previous single-image based calibration methods both on standard metrics as well as on this novel perceptual measure. Finally, we demonstrate the use of our calibration network for several applications, including virtual object insertion, image retrieval, and compositing.</p>
<h1>A Deep Perceptual Measure for Lens and Camera Calibration</h1>
<p>Image editing and compositing have become ubiquitous in entertainment, from digital art to AR and VR experiences. To produce beautiful composites, the camera needs to be geometrically calibrated, which can be tedious and requires a physical calibration target. In place of the traditional multi-image calibration process, we propose to infer the camera calibration parameters such as pitch, roll, field of view, and lens distortion directly from a single image using a deep convolutional neural network. We train this network using automatically generated samples from a large-scale panorama dataset, yielding competitive accuracy in terms of standard `2 error. However, we argue that minimizing such standard error metrics might not be optimal for many applications. In this work, we investigate human sensitivity to inaccuracies in geometric camera calibration. To this end, we conduct a large-scale human perception study where we ask participants to judge the realism of 3D objects composited with correct and biased camera calibration parameters. Based on this study, we develop a new perceptual measure for camera calibration and demonstrate that our deep calibration network outperforms previous single-image based calibration methods both on standard metrics as well as on this novel perceptual measure. Finally, we demonstrate the use of our calibration network for several applications, including virtual object insertion, image retrieval, and compositing.</p>
<img class="center-block" src="./images/teaser-new.jpg" width="800px">
</div>
</div>
Expand All @@ -70,11 +70,11 @@ <h1>A Perceptual Measure for Deep Single Image <br/> Camera and Lens Calibration
<div class="container">
<div class="row">
<div class="col-md-12">
<h2 class="section-heading">Paper (arXiv 2022 version)</h2>
<h2 class="section-heading">Paper (TPAMI 2023 version)</h2>

<a href="http://yannickhold.com">Yannick Hold-Geoffroy</a>, Dominique Piché-Meunier, <a href="https://research.adobe.com/person/kalyan-sunkavalli/">Kalyan Sunkavalli</a>, Jean-Charles Bazin, <a href="https://rameau-fr.github.io">François Rameau</a>, and <a href="http://vision.gel.ulaval.ca/~jflalonde/">Jean-François Lalonde</a> <br/>

A Perceptual Measure for Deep Single Image Camera and Lens Calibration <br/>
A Deep Perceptual Measure for Lens and Camera Calibration <br/>

[<a href="https://arxiv.org/abs/2208.12300">arXiv pre-print</a>]

Expand Down Expand Up @@ -107,7 +107,7 @@ <h4>Supplementary material</h4>
<div class="clearfix"></div>
<h2 class="section-heading">Live demo</h2>

<p>We have two different versions of our most recent (arXiv 2022) approach:
<p>We have two different versions of our most recent (TPAMI 2023) approach:
<ul>
<li><a href="http://rachmaninoff.gel.ulaval.ca:8004">With a ResNet backbone (as in the paper)</a></li>
<li><a href="http://rachmaninoff.gel.ulaval.ca:8005">With a more recent ConvNeXt backbone</a></li>
Expand Down Expand Up @@ -180,26 +180,25 @@ <h2 class="section-heading">Poster (CVPR 2018)</h2>
<h2 class="section-heading">Acknowledgements</h2>

<p>The authors gratefully acknowledge the following funding sources: </p>

<ul>
<li>NSERC Discovery GRANT RGPIN-2014-05314</li>
<li>Korea NRF grant NRF-2017R1C1B5077030</li>
<li>FRQ-NT Ph.D. scholarship to Yannick Hold-Geoffroy</li>
<li>NSERC USRA to Dominique Piché-Meunier</li>
<li>A generous donation from Adobe to J-F Lalonde and J-C Bazin</li>
<li>NVIDIA Corporation with the donation of the Tesla K40 and Titan X GPUs used for this research.</li>
<li>REPARTI Strategic Network</li>
</ul>
<ul>
<li>NSERC Discovery GRANT RGPIN-2014-05314</li>
<li>Korea NRF grant NRF-2017R1C1B5077030</li>
<li>FRQ-NT Ph.D. scholarship to Yannick Hold-Geoffroy</li>
<li>NSERC USRA to Dominique Piché-Meunier</li>
<li>A generous donation from Adobe to J-F Lalonde and J-C Bazin</li>
<li>NVIDIA Corporation with the donation of the Tesla K40 and Titan X GPUs used for this research.</li>
<li>REPARTI Strategic Network</li>
</ul>
</div>
</div>
<br>
<div class="row">
<div class="col-md-5">
<img src="./images/ulaval-logo-350.png" height="150px" alt="Ulaval logo">
</div>
<div class="col-md-5 col-md-offset-1" style="text-align: center;">
<img src="./images/adobe-logo.png" height="150px" alt="Adobe logo">
</div>
</div>
<div class="col-md-5 col-md-offset-1" style="text-align: center;">
<img src="./images/adobe-logo.png" height="150px" alt="Adobe logo">
</div>
</div>

</div>
Expand Down

0 comments on commit 6bfd54f

Please sign in to comment.