Jonathan Eisenmann's research while affiliated with Adobe Inc. and other places

Publications (10)

Preprint
We propose a method to extrapolate a 360{\deg} field of view from a single image that allows for user-controlled synthesis of the out-painted content. To do so, we propose improvements to an existing GAN-based in-painting architecture for out-painting panoramic image representation. Our method obtains state-of-the-art results and outperforms previo...
Article
High-quality denoising of Monte Carlo low-sample renderings remains a critical challenge for practical interactive ray tracing. We present a new learning-based denoiser that achieves state-of-the-art quality and runs at interactive rates. Our model processes individual path-traced samples with a lightweight neural network to extract per-pixel featu...
Chapter
Most 3D reconstruction methods may only recover scene properties up to a global scale ambiguity. We present a novel approach to single view metrology that can recover the absolute scale of a scene represented by 3D heights of objects or camera height above the ground as well as camera parameters of orientation and field of view, using just a monocu...
Conference Paper
Full-text available
Most 3D reconstruction methods may only recover scene properties up to a global scale ambiguity. We present a novel approach to single view metrology that can recover the \emph{absolute} scale of a scene represented by 3D heights of objects or camera height above the ground as well as camera parameters of orientation and field of view, using just a...
Preprint
Full-text available
Most 3D reconstruction methods may only recover scene properties up to a global scale ambiguity. We present a novel approach to single view metrology that can recover the absolute scale of a scene represented by 3D heights of objects or camera height above the ground as well as camera parameters of orientation and field of view, using just a monocu...
Preprint
We introduce UprightNet, a learning-based approach for estimating 2DoF camera orientation from a single RGB image of an indoor scene. Unlike recent methods that leverage deep learning to perform black-box regression from image to orientation parameters, we propose an end-to-end framework that incorporates explicit geometric reasoning. In particular...
Preprint
We present a neural network that predicts HDR outdoor illumination from a single LDR image. At the heart of our work is a method to accurately learn HDR lighting from LDR panoramas under any weather condition. We achieve this by training another CNN (on a combination of synthetic and real images) to take as input an LDR panorama, and regress the pa...
Article
Full-text available
Most current single image camera calibration methods rely on specific image features or user input, and cannot be applied to natural images captured in uncontrolled settings. We propose inferring directly camera calibration parameters from a single image using a deep convolutional neural network. This network is trained using automatically generate...

Citations

... For such synthetic content, there has been significant effort to improve sampling [Müller et al. 2019] or post-process denoising [Bako et al. 2017;Chaitanya et al. 2017;Işık et al. 2021] for Monte Carlo global illumination algorithms. Denoising methods fail to reproduce complex lighting effects (caustics, specular-diffusespecular paths) if they do not exist in the noisy input, while our method can reproduce them at interactive rates in high resolution (see Sec. 7.2.3). ...
... Günel et al. [19] introduce the IMDB-23K dataset by gathering publicly available celebrity images and their height information. Zhu et al. [74] use this dataset to learn to predict the height of people in images. Dey et al. [13] estimate the height of users in a photo collection by computing height differences between people in an image, creating a graph that links people across photos, and solving a maximum likelihood estimation problem. ...
... Learning-based calibration. Some of the most recent works [25][26][27][28] have leveraged the success of convolutional neural networks and have presented learning-based methods to estimate camera parameters. Some studies [27,28] have been able to estimate camera parameters from a single image and have reported a good calibration accuracy. ...
... Several deep learning-based approaches have recently been proposed to overcome such limitations [37,16,38]. These methods directly infer the camera parameters from an input image using semantic cues learned by deep neural networks. ...