1 code implementation • 26 Aug 2024 • Georg Bökman, Johan Edstedt, Michael Felsberg, Fredrik Kahl
We propose a way to train deep learning based keypoint descriptors that makes them approximately equivariant for locally affine transformations of the image plane.
no code implementations • 25 Aug 2024 • Ioannis Athanasiadis, Shashi Nagarajan, Fredrik Lindsten, Michael Felsberg
Variational Autoencoders (VAEs) are a popular framework for unsupervised learning and data generation.
no code implementations • 7 Jun 2024 • Arvi Jonnarth, Ola Johansson, Michael Felsberg
We bridge the sim-to-real gap through a semi-virtual environment, including a real robot and real-time aspects, while utilizing a simulated sensor and obstacles to enable environment randomization and automated episode resetting.
2 code implementations • 11 Apr 2024 • William Ljungbergh, Adam Tonderski, Joakim Johnander, Holger Caesar, Kalle Åström, Michael Felsberg, Christoffer Petersson
We present a versatile NeRF-based simulator for testing autonomous driving (AD) software systems, designed with a focus on sensor-realistic closed-loop evaluation and the creation of safety-critical scenarios.
1 code implementation • CVPR 2024 • Omkar Thawakar, Muzammal Naseer, Rao Muhammad Anwer, Salman Khan, Michael Felsberg, Mubarak Shah, Fahad Shahbaz Khan
Composed video retrieval (CoVR) is a challenging problem in computer vision which has recently highlighted the integration of modification text with visual queries for more sophisticated video search in large databases.
1 code implementation • 8 Mar 2024 • Yushan Zhang, Bastian Wandt, Maria Magnusson, Michael Felsberg
Aiming at improving accuracy while additionally providing an estimate for uncertainty, we propose DiffSF that combines transformer-based scene flow estimation with denoising diffusion models.
1 code implementation • 26 Feb 2024 • Omkar Thawakar, Ashmal Vayani, Salman Khan, Hisham Cholakal, Rao M. Anwer, Michael Felsberg, Tim Baldwin, Eric P. Xing, Fahad Shahbaz Khan
"Bigger the better" has been the predominant trend in recent Large Language Models (LLMs) development.
1 code implementation • 22 Feb 2024 • Muhammad Maaz, Hanoona Rasheed, Abdelrahman Shaker, Salman Khan, Hisham Cholakal, Rao M. Anwer, Tim Baldwin, Michael Felsberg, Fahad S. Khan
PALO offers visual reasoning capabilities in 10 major languages, including English, Chinese, Hindi, Spanish, French, Arabic, Bengali, Russian, Urdu, and Japanese, that span a total of ~5B people (65% of the world population).
no code implementations • 7 Jan 2024 • Pourya Shamsolmoali, Masoumeh Zareapoor, Eric Granger, Michael Felsberg
Kernel methods are employed to simplify computations by approximating softmax but often lead to performance drops compared to softmax attention.
1 code implementation • CVPR 2024 • Georg Bökman, Johan Edstedt, Michael Felsberg, Fredrik Kahl
Image keypoint descriptions that are discriminative and matchable over large changes in viewpoint are vital for 3D reconstruction.
no code implementations • 16 Oct 2023 • Hannah Helgesen, Michael Felsberg, Jan-Åke Larsson
The field of Quantum Machine Learning (QML) has emerged recently in the hopes of finding new machine learning protocols or exponential speedups for classical ones.
no code implementations • 15 Sep 2023 • Jie Zhao, Johan Edstedt, Michael Felsberg, Dong Wang, Huchuan Lu
Due to long-distance correlation and powerful pretrained models, transformer-based methods have initiated a breakthrough in visual object tracking performance.
no code implementations • 12 Sep 2023 • Qiyu Sun, Huilin Chen, Meng Zheng, Ziyan Wu, Michael Felsberg, Yang Tang
Domain generalized semantic segmentation (DGSS) is a critical yet challenging task, where the model is trained only on source data without access to any target data.
2 code implementations • 16 Aug 2023 • Johan Edstedt, Georg Bökman, Mårten Wadenbäck, Michael Felsberg
To train a descriptor, we maximize the mutual nearest neighbour objective over the keypoints with a separate network.
no code implementations • 4 Jul 2023 • Qiyu Sun, Pavlo Melnyk, Michael Felsberg, Yang Tang
Domain generalized semantic segmentation (DGSS) is an essential but highly challenging task, in which the model is trained only on source data and any target data is not available.
1 code implementation • 29 Jun 2023 • Arvi Jonnarth, Jie Zhao, Michael Felsberg
Coverage path planning (CPP) is the problem of finding a path that covers the entire free space of a confined area, with applications ranging from robotic lawn mowing to search-and-rescue.
1 code implementation • 7 Jun 2023 • Emanuel Sanchez Aimar, Nathaniel Helgesen, Yonghao Xu, Marco Kuhlmann, Michael Felsberg
Long-tailed semi-supervised learning (LTSSL) represents a practical scenario for semi-supervised applications, challenged by skewed labeled distributions that bias classifiers.
1 code implementation • NeurIPS 2023 • Yushan Zhang, Johan Edstedt, Bastian Wandt, Per-Erik Forssén, Maria Magnusson, Michael Felsberg
We tackle the task of scene flow estimation from point clouds.
no code implementations • 24 May 2023 • Pavlo Melnyk, Michael Felsberg, Mårten Wadenbäck, Andreas Robinson, Cuong Le
In this paper, we utilize hyperspheres and regular $n$-simplexes and propose an approach to learning deep features equivariant under the transformations of $n$D reflections and rotations, encompassed by the powerful group of O$(n)$.
1 code implementation • CVPR 2024 • Johan Edstedt, Qiyu Sun, Georg Bökman, Mårten Wadenbäck, Michael Felsberg
The aim is to learn a robust model, i. e., a model able to match under challenging real-world changes.
1 code implementation • 5 Apr 2023 • Arvi Jonnarth, Yushan Zhang, Michael Felsberg
Our work is based on two techniques for improving CAMs; importance sampling, which is a substitute for GAP, and the feature similarity loss, which utilizes a heuristic that object contours almost always align with color edges in images.
no code implementations • 25 Jan 2023 • Yushan Zhang, Andreas Robinson, Maria Magnusson, Michael Felsberg
A model to extract the combined information from optical flow and the image is proposed, which is then used as input to the target model and the decoder network.
no code implementations • 21 Jan 2023 • William Ljungbergh, Joakim Johnander, Christoffer Petersson, Michael Felsberg
Images fed to a deep neural network have in general undergone several handcrafted image signal processing (ISP) operations, all of which have been optimized to produce visually pleasing images.
no code implementations • 6 Dec 2022 • Karl Holmquist, Lena Klasén, Michael Felsberg
In this paper, we address the problem of how to model unlabeled classes while avoiding spurious feature clustering of future uncorrelated classes.
class-incremental learning Class-Incremental Semantic Segmentation +3
1 code implementation • CVPR 2024 • Pavlo Melnyk, Andreas Robinson, Michael Felsberg, Mårten Wadenbäck
In our approach, we perform TetraTransform--an equivariant embedding of the 3D input into 4D, constructed from the steerable neurons--and extract deeper O(3)-equivariant features using vector neurons.
1 code implementation • CVPR 2023 • Emanuel Sanchez Aimar, Arvi Jonnarth, Michael Felsberg, Marco Kuhlmann
We show how to properly define these distributions and combine the experts in order to achieve unbiased predictions, by proving that the ensemble is Fisher-consistent for minimizing the balanced error.
Long-tail Learning Long-tail Learning on CIFAR-10-LT (ρ=100) +1
1 code implementation • 24 Mar 2022 • Omkar Thawakar, Sanath Narayan, Jiale Cao, Hisham Cholakkal, Rao Muhammad Anwer, Muhammad Haris Khan, Salman Khan, Michael Felsberg, Fahad Shahbaz Khan
When using the ResNet50 backbone, our MS-STS achieves a mask AP of 50. 1 %, outperforming the best reported results in literature by 2. 7 % and by 4. 8 % at higher overlap threshold of AP_75, while being comparable in model size and speed on Youtube-VIS 2019 val.
1 code implementation • 23 Mar 2022 • Arvi Jonnarth, Michael Felsberg
Classification networks can be used to localize and segment objects in images by means of class activation maps (CAMs).
Weakly supervised Semantic Segmentation Weakly-Supervised Semantic Segmentation
no code implementations • 2 Mar 2022 • Oliver Stromann, Alireza Razavi, Michael Felsberg
In this work, we present a novel approach to learning an encoding of visual features into graph neural networks with the application on road network data.
1 code implementation • CVPR 2023 • Johan Edstedt, Ioannis Athanasiadis, Mårten Wadenbäck, Michael Felsberg
This changes with our novel dense method, which outperforms both dense and sparse methods on geometry estimation.
no code implementations • 20 Dec 2021 • Oliver Stromann, Alireza Razavi, Michael Felsberg
Road networks are the core infrastructure for connected and autonomous vehicles, but creating meaningful representations for machine learning applications is a challenging task.
no code implementations • 6 Dec 2021 • Sajid Javed, Martin Danelljan, Fahad Shahbaz Khan, Muhammad Haris Khan, Michael Felsberg, Jiri Matas
Accurate and robust visual object tracking is one of the most challenging and fundamental computer vision problems.
no code implementations • 6 Dec 2021 • Ankan Kumar Bhunia, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer, Fahad Shahbaz Khan, Jorma Laaksonen, Michael Felsberg
Creative sketch image generation is a challenging vision problem, where the task is to generate diverse, yet realistic creative sketches possessing the unseen composition of the visual-world objects.
1 code implementation • 7 Oct 2021 • Joakim Johnander, Johan Edstedt, Michael Felsberg, Fahad Shahbaz Khan, Martin Danelljan
Given the support set, our dense GP learns the mapping from local deep image features to mask values, capable of capturing complex appearance distributions.
Ranked #1 on Few-Shot Semantic Segmentation on COCO-20i (10-shot)
no code implementations • 29 Sep 2021 • Pavlo Melnyk, Michael Felsberg, Mårten Wadenbäck
Emerging from low-level vision theory, steerable filters found their counterpart in prior work on steerable convolutional neural networks equivariant to rigid transformations.
no code implementations • 29 Sep 2021 • Karl Holmquist, Michael Felsberg, Lena Klasen
In this paper we address the problem of how to model unlabeled classes to avoid unnecessary feature clustering of uncorrelated classes.
class-incremental learning Class-Incremental Semantic Segmentation +2
1 code implementation • 16 Jul 2021 • Zahra Gharaee, Shreyas Kowshik, Oliver Stromann, Michael Felsberg
We present a novel learning-based approach to graph representations of road networks employing state-of-the-art graph convolutional neural networks.
no code implementations • 15 Jun 2021 • Johan Edstedt, Amanda Berg, Michael Felsberg, Johan Karlsson, Francisca Benavente, Anette Novak, Gustav Grund Pihlgren
Automatically identifying harmful content in video is an important task with a wide range of applications.
1 code implementation • 2 Jun 2021 • Pavlo Melnyk, Michael Felsberg, Mårten Wadenbäck
In our work, we propose a steerable feed-forward learning-based approach that consists of neurons with spherical decision surfaces and operates on point clouds.
1 code implementation • International Conference on Patern Recognition (ICPR 2020) 2021 • Zahra Gharaee, Karl Holmquist, Linbo He, Michael Felsberg
We trained our system using both ground truth and estimated semantic segmentation input.
no code implementations • 30 Mar 2021 • Joakim Johnander, Johan Edstedt, Martin Danelljan, Michael Felsberg, Fahad Shahbaz Khan
Through the expressivity of the GP, our approach is capable of modeling complex appearance distributions in the deep feature space.
2 code implementations • 13 Feb 2021 • Abdelrahman Eldesokey, Michael Felsberg
Our proposed approach formulates the upsampling task as a sparse problem and employs the normalized convolutional neural networks to solve it.
no code implementations • 7 Dec 2020 • Joakim Johnander, Emil Brissman, Martin Danelljan, Michael Felsberg
Most existing approaches to video instance segmentation comprise multiple modules that are heuristically combined to produce the final output.
1 code implementation • ICCV 2021 • Pavlo Melnyk, Michael Felsberg, Mårten Wadenbäck
Our extension of the MLHP model, the multilayer geometric perceptron (MLGP), and its respective layer units, i. e., geometric neurons, are consistent with the 3D geometry and provide a geometric handle of the learned coefficients.
1 code implementation • CVPR 2020 • Abdelrahman Eldesokey, Michael Felsberg, Karl Holmquist, Mikael Persson
In this work, we thus focus on modeling the uncertainty of depth data in depth completion starting from the sparse noisy input all the way to the final prediction.
2 code implementations • ECCV 2020 • Goutam Bhat, Felix Järemo Lawin, Martin Danelljan, Andreas Robinson, Michael Felsberg, Luc van Gool, Radu Timofte
This allows us to achieve a rich internal representation of the target in the current frame, significantly increasing the segmentation accuracy of our approach.
2 code implementations • CVPR 2020 • Andreas Robinson, Felix Järemo Lawin, Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg
The target appearance model consists of a light-weight module, which is learned during the inference stage using fast optimization techniques to predict a coarse but robust target segmentation.
1 code implementation • 27 May 2019 • Amanda Berg, Jörgen Ahlberg, Michael Felsberg
In this work, we evaluate the effects of anomaly contaminations in the training data on state-of-the-art GAN-based anomaly detection methods.
no code implementations • 18 Apr 2019 • Andreas Robinson, Felix Järemo Lawin, Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg
We propose a novel approach, based on a dedicated target appearance model that is exclusively learned online to discriminate between the target and background image regions.
1 code implementation • CVPR 2019 • Joakim Johnander, Martin Danelljan, Emil Brissman, Fahad Shahbaz Khan, Michael Felsberg
One of the fundamental challenges in video object segmentation is to find an effective representation of the target and background appearance.
4 code implementations • CVPR 2019 • Martin Danelljan, Goutam Bhat, Fahad Shahbaz Khan, Michael Felsberg
We argue that this approach is fundamentally limited since target estimation is a complex task, requiring high-level knowledge about the object.
Ranked #7 on Object Tracking on FE108
1 code implementation • 5 Nov 2018 • Abdelrahman Eldesokey, Michael Felsberg, Fahad Shahbaz Khan
In this paper, we propose an algebraically-constrained normalized convolution layer for CNNs with highly sparse input that has a smaller number of network parameters compared to related work.
Ranked #7 on Depth Completion on KITTI Depth Completion
1 code implementation • 30 May 2018 • Abdelrahman Eldesokey, Michael Felsberg, Fahad Shahbaz Khan
To tackle this challenging problem, we introduce an algebraically-constrained convolution layer for CNNs with sparse input and demonstrate its capabilities for the scene depth completion task.
no code implementations • ECCV 2018 • Goutam Bhat, Joakim Johnander, Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg
In the field of generic object tracking numerous attempts have been made to exploit deep features.
1 code implementation • CVPR 2018 • Felix Järemo Lawin, Martin Danelljan, Fahad Shahbaz Khan, Per-Erik Forssén, Michael Felsberg
Contrary to previous works, we model the underlying structure of the scene as a latent probability distribution, and thereby induce invariance to point set density changes.
no code implementations • 9 Jun 2017 • Joakim Johnander, Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg
Generally, DCF based trackers learn a rigid appearance model of the target.
1 code implementation • 9 May 2017 • Felix Järemo Lawin, Martin Danelljan, Patrik Tosteberg, Goutam Bhat, Fahad Shahbaz Khan, Michael Felsberg
Recent attempts, based on 3D deep learning approaches (3D-CNNs), have achieved below-expected results.
Ranked #15 on Semantic Segmentation on Semantic3D
no code implementations • 20 Dec 2016 • Susanna Gladh, Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg
To the best of our knowledge, we are the first to propose fusing appearance information with deep motion features for visual tracking.
no code implementations • 14 Dec 2016 • Fahad Shahbaz Khan, Joost Van de Weijer, Rao Muhammad Anwer, Andrew D. Bagdanov, Michael Felsberg, Jorma Laaksonen
Most approaches to human attribute and action recognition in still images are based on image representation in which multi-scale local features are pooled across scale into a single, scale-invariant encoding.
5 code implementations • CVPR 2017 • Martin Danelljan, Goutam Bhat, Fahad Shahbaz Khan, Michael Felsberg
Moreover, our fast variant, using hand-crafted features, operates at 60 Hz on a single CPU, while obtaining 65. 0% AUC on OTB-2015.
Ranked #13 on Visual Object Tracking on VOT2017/18
no code implementations • 20 Sep 2016 • Martin Danelljan, Gustav Häger, Fahad Shahbaz Khan, Michael Felsberg
Compared to the standard exhaustive scale search, our approach achieves a gain of 2. 5% in average overlap precision on the OTB dataset.
no code implementations • CVPR 2016 • Martin Danelljan, Gustav Häger, Fahad Shahbaz Khan, Michael Felsberg
We propose a novel generic approach for alleviating the problem of corrupted training samples in tracking-by-detection frameworks.
no code implementations • ICCV 2015 • Martin Danelljan, Gustav Häger, Fahad Shahbaz Khan, Michael Felsberg
These methods utilize a periodic assumption of the training samples to efficiently learn a classifier on all patches in the target neighborhood.
1 code implementation • 12 Aug 2016 • Martin Danelljan, Andreas Robinson, Fahad Shahbaz Khan, Michael Felsberg
We also demonstrate the effectiveness of our learning formulation in extensive feature point tracking experiments.
no code implementations • CVPR 2016 • Martin Danelljan, Giulia Meneghetti, Fahad Shahbaz Khan, Michael Felsberg
On the Stanford Lounge dataset, our approach achieves a relative reduction of the failure rate by 78% compared to the baseline.
no code implementations • 29 Jan 2016 • Erik Jonsson, Michael Felsberg
A robust mean value is often a good alternative to the standard mean value when dealing with data containing many outliers.
no code implementations • CVPR 2014 • Martin Danelljan, Fahad Shahbaz Khan, Michael Felsberg, Joost Van de Weijer
This paper investigates the contribution of color in a tracking-by-detection framework.