1 code implementation • 7 Aug 2023 • Artem Moskalev, Anna Sepliarskaia, Erik J. Bekkers, Arnold Smeulders
We demonstrate that even when a network learns to correctly classify samples on a group orbit, the underlying decision-making in such a model does not attain genuine invariance.
no code implementations • 12 Jan 2023 • Ivan Sosnovik, Artem Moskalev, Cees Kaandorp, Arnold Smeulders
Video summarization aims at choosing parts of a video that narrate a story as close as possible to the original one.
1 code implementation • 9 Oct 2022 • Artem Moskalev, Anna Sepliarskaia, Ivan Sosnovik, Arnold Smeulders
Symmetries built into a neural network have appeared to be very beneficial for a wide range of tasks as it saves the data to learn them.
1 code implementation • 31 May 2022 • Artem Moskalev, Ivan Sosnovik, Volker Fischer, Arnold Smeulders
The views are ordered in pairs, such that they are either positive, encoding different views of the same object, or negative, corresponding to views of different objects.
no code implementations • 1 Dec 2021 • Mert Kilickaya, Arnold Smeulders
iii) We evaluate Align-Former on HICO-DET [5] and V-COCO [13], and show that Align-Former outperforms existing image-level supervised HO-I detectors by a large margin (4. 71% mAP improvement from 16. 14% to 20. 85% on HICO-DET [5]).
1 code implementation • 18 Nov 2021 • Sadaf Gulshad, Ivan Sosnovik, Arnold Smeulders
To demonstrate that wiggling the weights consistently improves classification, we choose a standard network and modify it to a transform-augmented network.
no code implementations • ICLR 2019 • Jan Jetze Beitler, Ivan Sosnovik, Arnold Smeulders
We consider the problem of information compression from high dimensional data.
no code implementations • 11 Aug 2021 • Artem Moskalev, Ivan Sosnovik, Arnold Smeulders
Tracking multiple objects individually differs from tracking groups of related objects.
no code implementations • ICCVW 2021 • Ivan Sosnovik, Artem Moskalev, Arnold Smeulders
We aim for accurate scale-equivariant convolutional neural networks (SE-CNNs) applicable for problems where high granularity of scale and small kernel sizes are required.
no code implementations • 20 Jul 2021 • Sadaf Gulshad, Ivan Sosnovik, Arnold Smeulders
We focus on building robustness in the convolutions of neural visual classifiers, especially against natural perturbations like elastic deformations, occlusions and Gaussian noise.
1 code implementation • 4 Jun 2021 • Ivan Sosnovik, Artem Moskalev, Arnold Smeulders
In recent work scale equivariance was added to convolutional neural networks.
no code implementations • 21 Mar 2021 • Sadaf Gulshad, Arnold Smeulders
For Cifar-10 and STL-10 natural perturbed training even improves the accuracy for clean data and reaches the state of the art performance.
no code implementations • 17 Oct 2020 • Mert Kilickaya, Noureldien Hussein, Efstratios Gavves, Arnold Smeulders
Our experiments show that SSC leads to an important increase in interaction recognition performance, while using much fewer parameters.
no code implementations • 3 Oct 2020 • Sadaf Gulshad, Jan Hendrik Metzen, Arnold Smeulders
In this paper we aim to explore the general robustness of neural network classifiers by utilizing adversarial as well as natural perturbations.
1 code implementation • 17 Jul 2020 • Ivan Sosnovik, Artem Moskalev, Arnold Smeulders
We develop the theory for scale-equivariant Siamese trackers, and provide a simple recipe for how to make a wide range of existing trackers scale-equivariant.
Ranked #1 on Visual Object Tracking on OTB-2013
no code implementations • 10 Jun 2020 • Mert Kilickaya, Arnold Smeulders
To that end, in this paper, we propose to diagnose rarity in HOI detection.
1 code implementation • 27 Jan 2020 • Sadaf Gulshad, Arnold Smeulders
Hence, inspired by the way of human explanations in this paper we provide attribute-based and example-based explanations.
1 code implementation • 15 Oct 2019 • Sadaf Gulshad, Zeynep Akata, Jan Hendrik Metzen, Arnold Smeulders
We study the changes in attributes for clean as well as adversarial images in both standard and adversarially robust networks.
1 code implementation • ICLR 2020 • Ivan Sosnovik, Michał Szmaja, Arnold Smeulders
The effectiveness of Convolutional Neural Networks (CNNs) has been substantially attributed to their built-in property of translation equivariance.
Ranked #33 on Image Classification on STL-10
1 code implementation • 17 Apr 2019 • Sadaf Gulshad, Jan Hendrik Metzen, Arnold Smeulders, Zeynep Akata
Deep computer vision systems being vulnerable to imperceptible and carefully crafted noise have raised questions regarding the robustness of their decisions.
no code implementations • ECCV 2018 • Jack Valmadre, Luca Bertinetto, João F. Henriques, Ran Tao, Andrea Vedaldi, Arnold Smeulders, Philip Torr, Efstratios Gavves
We introduce the OxUvA dataset and benchmark for evaluating single-object tracking algorithms.
2 code implementations • ICLR 2018 • Jörn-Henrik Jacobsen, Arnold Smeulders, Edouard Oyallon
An analysis of i-RevNets learned representations suggests an alternative explanation for the success of deep networks by a progressive contraction and linear separation with depth.
no code implementations • NeurIPS 2009 • Sennay Ghebreab, Steven Scholte, Victor Lamme, Arnold Smeulders
From these neural measurements and the contrast statistics of the natural image stimuli, we derive an across subject Weibull response model.
no code implementations • NeurIPS 2007 • Gertjan Burghouts, Arnold Smeulders, Jan-Mark Geusebroek
This fundamental insight opens new directions in the assessment of feature similarity, with projected improvements in object and scene recognition algorithms.
no code implementations • NeurIPS 2007 • Sennay Ghebreab, Arnold Smeulders, Pieter Adriaans
We propose a method for reconstruction of human brain states directly from functional neuroimaging data.