The Sky’s the Limit: Relightable Outdoor Scenes via a Sky-Pixel Constrained Illumination Prior and Outside-In Visibility
Use sky pixels as direct observations to constrain statistical illumination model + differentiable shadows!
You Only Look for a Symbol Once: An Object Detector for Symbols and Regions in Documents
YOLO adapted for symbols and regions in map-style data with no meaningful boundaries.
HexNet: An Orientation-aware Deep Learning Framework for Omni-directional Input
CNN for spherical images on a planar hexagonal grid.
Self-Supervised Relative Pose With Homography Model-Fitting in the Loop
A compact, neural model for representing natural illumination
Rotation-Equivariant Conditional Spherical Neural Fields for Learning a Natural Illumination Prior
A compact, neural model for representing natural illumination
NeRF for Outdoor Scene Relighting
Learn a neural representation of geometry, albedo and shadows from which we can edit both lighting and viewpoint.
Outdoor inverse rendering from a single image using multiview self-supervision
We learn single image inverse rendering of surface normal, albedo, shadow maps and lighting from unstructured outdoor image …
Unravelling the long-term, locally heterogenous response of Greenland glaciers observed in archival photography
SfM on Greenland glacier photographs from 1930s to present.
Structure-from-motion with varying principal point
SfM for image sequences with varying principal point, e.g. from glass plate cameras.
SurfaceView: Seamless and tile-based orthomosaics using millions of street-level images from vehicle-mounted cameras
Merge millions of streetview images into seamless virtual top down view images.
Uncalibrated, Two Source Photo-Polarimetric Stereo
Photo-polarimetric shape estimation, under varying, but unknown, illumination.
Evaluation of structure-from-motion for analysis of small-scale glacier dynamics
We propose a measurement procedure for marine-terminating glaciers using structure-from-motion.
Shape from semantic segmentation via the geometric Renyi divergence
A soft loss function for measuring consistency between projected 3D model and 2D semantic segmentation.
Reconstructing Creative Lego Models
From a physical lego model to a semantic reconstruction in terms of relative brick placements.
Self-supervised Outdoor Scene Relighting
Train a neural renderer to use inverse rendered output for photorealistic relighting without requiring any supervision.
Look Ma, no landmarks! – Unsupervised, Model-based Dense Face Alignment
Estimate pixel-wise correspondence to a 3DMM, solve for all 3DMM parameters with linear least squares, no supervision required!
Least squares surface reconstruction on arbitrary domains
Arbitrary foreground mask -> custom 2D Savitzky-Golay filters using pixel KNN -> surface integration with linear least squares.
A Morphable Face Albedo Model
We capture a dataset of 50 scans, combine them with a publicly available albedo dataset and build the first morphable face albedo …
3D Morphable Face Models - Past, Present and Future
A detailed survey of 3D Morphable Face Models over the 20 years since they were first proposed.
Towards a complete 3D morphable model of the human head
We present the most complete 3DMM of the human head to date that includes face, cranium, ears, eyes, teeth and tongue.
A Hierarchical Loss for Semantic Segmentation
We exploit knowledge of class hierarchies to aid the training of semantic segmentation convolutional neural networks.
Statistical Modeling of Craniofacial Shape and Texture
We present a fully-automatic statistical 3D shape modeling approach and apply it to a large dataset of 3D images.
Depth estimation meets inverse rendering for single image novel view synthesis
We show how to combine state-of-the-art deep learning based methods for single image depth estimation and inverse rendering.
A Shadow Constrained Conditional Generative Adversarial Net for SRTM Data Restoration
We propose a shadow constrained conditional generative adversarial network for filling voids in incomplete SRTM data.
Height-from-Polarisation with Unknown Lighting or Albedo
We present a method for estimating surface height directly from a single polarisation image simply by solving a large, sparse system of …
Augmenting a 3D morphable model of the human head with high resolution ears
We present a parts-based 3D Morphable Model (3DMM) of the full human head, with particular emphasis on modelling the complex shape of …
Orientation-aware semantic segmentation on icosahedron spheres
We address semantic segmentation on omnidirectional images, to leverage a holistic understanding of the surrounding scene for …
3D Surface Texture Analysis of High Resolution Normal Fields for Facial Skin Condition Assessment
Investigates the use of a lightstage to capture high-resolution, 3D facial surface textures and proposes novel methods to use the data …
Decomposing multispectral face images into diffuse and specular shading and biophysical parameters
We propose a novel biophysical and dichromatic reflectance model that efficiently characterises spectral skin reflectance.
BioFaceNet: Deep Biophysical Face Image Interpretation
We propose a deep CNN that learns to decompose a single face image into biophysical parameters maps, diffuse and specular shading maps …
Filling Voids in Elevation Models Using a Shadow-Constrained Convolutional Neural Network
We explore the use of convolutional neural networks (CNNs) for filling voids in digital elevation models (DEM).
What Does 2D Geometric Information Really Tell Us About 3D Face Shape?
We show that uncalibrated, monocular face shape-from-correspondence (e.g. landmarks) with a 3D morphable model is highly ambiguous.
InverseRenderNet: Learning single image inverse rendering
We show how to train a fully convolutional neural network to perform inverse rendering from a single, uncontrolled image.
Depth from a polarisation + RGB stereo pair
We propose a hybrid depth imaging system in which a polarisation camera is augmented by a second image from a standard digital camera.
Combining 3D Morphable Models: A Large scale Face-and-Head Model
We show how to combine two or more 3DMMs that are built using different templates and have different representation capabilities. We …
Principal Geodesic Analysis in the Space of Discrete Shells
We propose a novel, nonlinear, rigid body motion invariant Principal Geodesic Analysis (PGA) that allows us to compress large shape …
Linear Differential Constraints for Photo-polarimetric Height Estimation
We present a differential approach to photopolarimetric shape estimation.
Symmetry-Aware Mesh Segmentation into Uniform Overlapping Patches
We present intrinsic methods to address the fundamental problem of segmenting a mesh into a specified number of patches with a uniform …
BRISKS: Binary Features for Spherical Images on a Geodesic Grid
We develop an interest point detector and binary feature descriptor for spherical images using a geodesic grid representation.
Structure-from-motion in Spherical Video using the von Mises-Fisher Distribution
In this paper we present a complete pipeline for computing structure-from-motion from sequences of spherical images.
Linear depth estimation from an uncalibrated, monocular polarisation image
We present a method for estimating surface height directly from a single polarisation image simply by solving a large, sparse system of …
Modelling the perceptual similarity of facial expressions from image statistics and neural responses
We measured the perceptual similarity between expressions and investigated how this is reflected in image measures and in the neural …
Manifold-based Constraints for Operations in Face Space
We constrain faces to points on a manifold of constant distinctiveness within the parameter space of a linear statistical model.
Height from Photometric Ratio with Model-based Light Source Selection
We present a photometric stereo algorithm for estimating surface height directly.
Statistical 3D Face Shape Estimation from Occluding Contours
We study the problem of 3D face shape approximation from occluding contours using a linear regression process.
Shell PCA: statistical shape modelling in shell space
We describe how to perform Principal Components Analysis in shell space.
Example-Based Modeling of Facial Texture from Deficient Data
Reconstruct ear-to-ear, high-quality texture from one or more partial views of a face with possibly poor resolution and noise.
Inverse rendering of faces with a 3D morphable model
Efficient 3DMM fitting with spherical harmonic lighting and specular reflectance.
