SphereNet: Learning Spherical Representations for Detection and Classification in Omnidirectional Images

By Benjamin Coors, Alexandru Paul Condurache, Andreas Geiger. 2018

436 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

WoodScape: A Multi-Task, Multi-Camera Fisheye Dataset for Autonomous Driving (2019, methodology): “Some examples of the few datasets that have fisheye are: Visual SLAM ground truth for indoor scenes with omnidirectional cameras in [7], SphereNet [9] containing 1200 labelled images of parked cars using 360° cameras (not strictly fisheye) and, in automotive, the Oxford Robotcar”
Spherical CNNs on Unstructured Grids (2019, methodology): “We benchmark our model against two other implementations of spherical CNNs: a rotational-invariant model by Cohen et al. (2018) and an orientable model by Coors et al. (2018).”
Emergence of exploratory look-around behaviors through active observation completion (2019, background): “Our approach was also affected by a well-known limitation associated with rectangular representations of spherical environments (44) where information at the poles are oversampled compared with the central elevations, resulting in redundant information across different azimuths at the poles.”
DeepSphere: Efficient spherical Convolutional Neural Network with HEALPix sampling for cosmological applications (2018, methodology): “A ﬁrst approach, explored by [20] for molecular modeling and [21, 22] for omnidirectional imaging, is to use a 2D CNN on a discretisation of the sphere that is a grid, such as the equirectangular projection (Figure 6, left panel), or a set of grids, such as the cubed-sphere deﬁned by [47] (Figure…”
Improving Data‐Driven Global Weather Prediction Using Deep Convolutional Neural Networks on a Cubed Sphere (2020, methodology): “[10] show that their method performs marginally better than a CNN on the cubed sphere.”
DeepSphere: a graph-based spherical CNN (2020, methodology): “Exploiting the locally Euclidean property of the sphere, early attempts used standard 2D convolutions on a grid sampling of the sphere (Boomsma & Frellsen, 2017; Su & Grauman, 2017; Coors et al., 2018).”
Kernel Transformer Networks for Compact Spherical Convolution (2018, methodology): “Compared with Spherical U-Net [46] and SphereNet [12], KTN is much more data efﬁcient because it does not require any annotated 360 ◦ images for training, and it is more accurate because it avoids their feature interpolation assumption.”
Encoder–decoder network for pixel‐level road crack detection in black‐box images (2019, background): “Third, the images are distorted based on the location of the black-box camera on the vehicle, unlike the situation in which a camera is mounted on a vehicle and is taking orthogonal images (Coors, Condurache, & Geiger, 2018; Rezaei, Sarrafzadeh, & Shanbehzadeh, 2013).”
Geometric Structure Based and Regularized Depth Estimation From 360 Indoor Imagery (2020, methodology): “Following [3], we also replace the planer convolution in FCRN with a spherical convolution and denote this baseline as spherical FCRN. 3 Result.”