OmniCV is a workshop in tandem with IEEE Computer Society Conference on Computer Vision and Pattern Recognition. The workshop seeks to mend the divide between research and application of omnidirectional vision technologies by offering a venue for research targeted towards realizing these ideas for commercial or societal benefit.
Friday, July 25th, 2021 (Virtually)
9am- 5pm US Central Time
POSTER SESSION: 1:15pm to 2pm
https://sites.google.com/view/omnicv2021/home
Presenter: Julie Buquet / PhD Candidate, Physics Department, Laval University -Intern Imaging (Immervision)
Evaluating the Impact of Wide-Angle Lens Distortion on Learning-based Depth Estimation
With the global increasing interest in wide angle imaging systems, we provide in this demo a study of the influence of their strong apparent distortion on learning-based computer vision applications. We use imaging systems with nonlinear distortion to control the pixel density in a region of interest on the image. Taking single image depth estimation as a case study, we show that using such property leads to local better neural networks accuracy compared to linear wide-angle imaging systems.
Julie Buquet is a PhD candidate at Laval University, QC, Canada. After getting her Master degree in Optics applied to computer vision at Institut d’Optique Graduate School, Paris, France, she joined Laval University on January 2020. Under the co-supervision of Simon Thibault, SPIE and OSA Fellow (Laboratory of Optics, Photonics and Laser, QC, Canada) and Jean-François Lalonde (Computer Vision and Systems Laboratory) she now specializes in wide-angle imaging systems applied to supervised learning based approaches. As an intern at Immervision, Julie benefits from their expertise in smart wide-angle imaging systems to complete her research, and to apply the results to industrial applications
Don’t Wait — Register Now!
Register Now to watch Julie’s demo about evaluating the impact of wide-angle lens distortion on learning-based depth estimation and attend the online Q&A session.