Search Results for author: Ondrej Chum

Found 22 papers, 9 papers with code

Local Features and Visual Words Emerge in Activations

1 code implementation15 May 2019 Oriane Siméoni, Yannis Avrithis, Ondrej Chum

We propose a novel method of deep spatial matching (DSM) for image retrieval.

Image Retrieval

Label Propagation for Deep Semi-supervised Learning

1 code implementation CVPR 2019 Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Ondrej Chum

In this work, we employ a transductive label propagation method that is based on the manifold assumption to make predictions on the entire dataset and use these predictions to generate pseudo-labels for the unlabeled data and train a deep neural network.

Local Orthogonal-Group Testing

no code implementations ECCV 2018 Ahmet Iscen, Ondrej Chum

This work addresses approximate nearest neighbor search applied in the domain of large-scale image retrieval.

Image Retrieval

Hybrid Diffusion: Spectral-Temporal Graph Filtering for Manifold Ranking

no code implementations23 Jul 2018 Ahmet Iscen, Yannis Avrithis, Giorgos Tolias, Teddy Furon, Ondrej Chum

State of the art image retrieval performance is achieved with CNN features and manifold ranking using a k-NN similarity graph that is pre-computed off-line.

Image Retrieval

Rectification from Radially-Distorted Scales

1 code implementation16 Jul 2018 James Pritts, Zuzana Kukelova, Viktor Larsson, Ondrej Chum

This paper introduces the first minimal solvers that jointly estimate lens distortion and affine rectification from repetitions of rigidly transformed coplanar local features.

Rectification

Mining on Manifolds: Metric Learning without Labels

1 code implementation CVPR 2018 Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Ondrej Chum

Positive examples are distant points on a single manifold, while negative examples are nearby points on different manifolds.

General Classification Metric Learning

Radially-Distorted Conjugate Translations

1 code implementation CVPR 2018 James Pritts, Zuzana Kukelova, Viktor Larsson, Ondrej Chum

The solvers are derived from constraints induced by the conjugate translations of an imaged scene plane, which are integrated with the division model for radial lens distortion.

Rectification

Coplanar Repeats by Energy Minimization

no code implementations26 Nov 2017 James Pritts, Denys Rozumnyi, M. Pawan Kumar, Ondrej Chum

This paper proposes an automated method to detect, group and rectify arbitrarily-arranged coplanar repeated elements via energy minimization.

Unsupervised object discovery for instance recognition

no code implementations14 Sep 2017 Oriane Siméoni, Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Ondrej Chum

Eliminating the impact of the clutter on the image descriptor increases the chance of retrieving relevant images and prevents topic drift due to actually retrieving the clutter in the case of query expansion.

Image Retrieval Object Discovery

Multiple-Kernel Local-Patch Descriptor

no code implementations25 Jul 2017 Arun Mukundan, Giorgos Tolias, Ondrej Chum

We propose a multiple-kernel local-patch descriptor based on efficient match kernels of patch gradients.

Panorama to panorama matching for location recognition

no code implementations21 Apr 2017 Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Teddy Furon, Ondrej Chum

Location recognition is commonly treated as visual instance retrieval on "street view" imagery.

Efficient Diffusion on Region Manifolds: Recovering Small Objects with Compact CNN Representations

3 code implementations CVPR 2017 Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Teddy Furon, Ondrej Chum

The diffusion is carried out on descriptors of overlapping image regions rather than on a global image descriptor like in previous approaches.

Image Retrieval

In the Saddle: Chasing Fast and Repeatable Features

1 code implementation24 Aug 2016 Javier Aldana-Iuit, Dmytro Mishkin, Ondrej Chum, Jiri Matas

A novel similarity-covariant feature detector that extracts points whose neighbourhoods, when treated as a 3D intensity surface, have a saddle-like intensity profile.

Camera Elevation Estimation from a Single Mountain Landscape Photograph

no code implementations12 Jul 2016 Martin Cadik, Jan Vasicek, Michal Hradis, Filip Radenovic, Ondrej Chum

This work addresses the problem of camera elevation estimation from a single photograph in an outdoor environment.

From Dusk Till Dawn: Modeling in the Dark

no code implementations CVPR 2016 Filip Radenovic, Johannes L. Schonberger, Dinghuang Ji, Jan-Michael Frahm, Ondrej Chum, Jiri Matas

We present an algorithm that leverages the appearance variety to obtain more complete and accurate scene geometry along with consistent multi-illumination appearance information.

Low Dimensional Explicit Feature Maps

no code implementations ICCV 2015 Ondrej Chum

Approximating non-linear kernels by finite-dimensional feature maps is a popular approach for speeding up training and evaluation of support vector machines or to encode information into efficient match kernels.

Multiple Measurements and Joint Dimensionality Reduction for Large Scale Image Search with Short Vectors - Extended Version

no code implementations13 Apr 2015 Filip Radenovic, Herve Jegou, Ondrej Chum

This paper addresses the construction of a short-vector (128D) image representation for large-scale image and particular object retrieval.

Dimensionality Reduction Image Retrieval

Detection, Rectification and Segmentation of Coplanar Repeated Patterns

no code implementations CVPR 2014 James Pritts, Ondrej Chum, Jiri Matas

This paper presents a novel and general method for the detection, rectification and segmentation of imaged coplanar repeated patterns.

Rectification

Cannot find the paper you are looking for? You can Submit a new open access paper.