no code implementations • 26 Apr 2024 • Jyri Maanpää, Julius Pesonen, Heikki Hyyti, Iaroslav Melekhov, Juho Kannala, Petri Manninen, Antero Kukko, Juha Hyyppä
We trained a convolutional neural network to predict pixelwise grip values from fused RGB camera, thermal camera, and LiDAR reflectance images, based on weakly supervised ground truth from an optical road weather sensor.
1 code implementation • 26 Mar 2024 • Matias Turkulainen, Xuqian Ren, Iaroslav Melekhov, Otto Seiskari, Esa Rahtu, Juho Kannala
3D Gaussian splatting, a novel differentiable rendering technique, has achieved state-of-the-art novel view synthesis results with high rendering speeds and relatively low training times.
1 code implementation • 20 Mar 2024 • Otto Seiskari, Jerry Ylilammi, Valtteri Kaatrasalo, Pekka Rantalankila, Matias Turkulainen, Juho Kannala, Esa Rahtu, Arno Solin
High-quality scene reconstruction and novel view synthesis based on Gaussian Splatting (3DGS) typically require steady, high-quality photographs, often impractical to capture with handheld cameras.
no code implementations • 5 Nov 2023 • Xuqian Ren, Wenjia Wang, Dingding Cai, Tuuli Tuominen, Juho Kannala, Esa Rahtu
Metaverse technologies demand accurate, real-time, and immersive modeling on consumer-grade hardware for both non-human perception (e. g., drone/robot/autonomous car navigation) and immersive technologies like AR/VR, requiring both structural accuracy and photorealism.
1 code implementation • 3 Nov 2023 • Wenshuai Zhao, Yi Zhao, Zhiyuan Li, Juho Kannala, Joni Pajarinen
However, with function approximation optimism can amplify overestimation and thus fail on complex tasks.
no code implementations • 23 Oct 2023 • Maximilian Krahn, Michelle Sasdelli, Fengyi Yang, Vladislav Golyanik, Juho Kannala, Tat-Jun Chin, Tolga Birdal
We present, QP-SBGD, a novel layer-wise stochastic optimiser tailored towards training neural networks with binary weights, known as binary neural networks (BNNs), on quantum hardware.
1 code implementation • 21 Jun 2023 • Shuzhe Wang, Juho Kannala, Daniel Barath
Matching 2D keypoints in an image to a sparse 3D point cloud of the scene without requiring visual descriptors has garnered increased interest due to its low memory requirements, inherent privacy preservation, and reduced need for expensive 3D model maintenance compared to visual descriptor-based methods.
1 code implementation • 15 Jun 2023 • Yi Zhao, Wenshuai Zhao, Rinu Boney, Juho Kannala, Joni Pajarinen
This applies when using pure planning with a dynamics model conditioned on the representation, but, also when utilizing the representation as policy and value function features in model-free RL.
no code implementations • 5 May 2023 • Shuzhe Wang, Zakaria Laskar, Iaroslav Melekhov, Xiaotian Li, Yi Zhao, Giorgos Tolias, Juho Kannala
In this work, we present a new hierarchical scene coordinate network to predict pixel scene coordinates in a coarse-to-fine manner from a single RGB image.
1 code implementation • 20 Feb 2023 • Masud Fahim, Ilona Söchting, Luca Ferranti, Juho Kannala, Jani Boutellier
Usually the query images have been acquired with a camera that differs from the imaging hardware used to collect the 3D database; consequently, it is hard to acquire accurate ground truth poses between query images and the 3D database.
no code implementations • 3 Jan 2023 • Janne Mustaniemi, Juho Kannala, Esa Rahtu, Li Liu, Janne Heikkilä
Various datasets have been proposed for simultaneous localization and mapping (SLAM) and related problems.
no code implementations • ICCV 2023 • Shuzhe Wang, Juho Kannala, Marc Pollefeys, Daniel Barath
We propose a new method, named curvature similarity extractor (CSE), for improving local feature matching across images.
1 code implementation • 27 Dec 2022 • Yingtian Zou, Vikas Verma, Sarthak Mittal, Wai Hoh Tang, Hieu Pham, Juho Kannala, Yoshua Bengio, Arno Solin, Kenji Kawaguchi
Mixup is a popular data augmentation technique for training deep neural networks where additional samples are generated by linearly interpolating pairs of inputs and their labels.
1 code implementation • 28 Nov 2022 • Hao Dong, Xianjing Zhang, Jintao Xu, Rui Ai, Weihao Gu, Huimin Lu, Juho Kannala, Xieyuanli Chen
However, current works are based on raw data or network feature-level fusion and only consider short-range HD map generation, limiting their deployment to realistic autonomous driving applications.
no code implementations • 1 Nov 2022 • Andrea Pilzer, Yuxin Hou, Niki Loppi, Arno Solin, Juho Kannala
We introduce visual hints expansion for guiding stereo matching to improve generalization.
2 code implementations • 25 Oct 2022 • Yi Zhao, Rinu Boney, Alexander Ilin, Juho Kannala, Joni Pajarinen
Offline reinforcement learning, by learning from a fixed dataset, makes it possible to learn agent behaviors without interacting with the environment.
1 code implementation • 4 Oct 2022 • Kalle Kujanpää, Amin Babadi, Yi Zhao, Juho Kannala, Alexander Ilin, Joni Pajarinen
To address this problem, we propose Continuous Monte Carlo Graph Search (CMCGS), an extension of MCTS to online planning in environments with continuous state and action spaces.
1 code implementation • 16 Aug 2022 • Subhankar Roy, Martin Trapp, Andrea Pilzer, Juho Kannala, Nicu Sebe, Elisa Ricci, Arno Solin
Source-free domain adaptation (SFDA) aims to adapt a classifier to an unlabelled target data set by only using a pre-trained source model.
1 code implementation • 14 Aug 2022 • Siyan Dong, Shuzhe Wang, Yixin Zhuang, Juho Kannala, Marc Pollefeys, Baoquan Chen
Visual (re)localization addresses the problem of estimating the 6-DoF (Degree of Freedom) camera pose of a query image captured in a known scene, which is a key building block of many computer vision and robotics applications.
no code implementations • 17 Jun 2022 • Ari Heljakka, Martin Trapp, Juho Kannala, Arno Solin
This observed 'predictive' multiplicity (PM) also implies elusive differences in the internals of the models, their 'representational' multiplicity (RM).
no code implementations • 15 Oct 2021 • Pauliina Paavilainen, Saad Ullah Akram, Juho Kannala
Medical image translation has the potential to reduce the imaging workload, by removing the need to capture some sequences, and to reduce the annotation burden for developing machine learning methods.
no code implementations • 10 Oct 2021 • Iaroslav Melekhov, Zakaria Laskar, Xiaotian Li, Shuzhe Wang, Juho Kannala
Fully-supervised CNN-based approaches for learning local image descriptors have shown remarkable results in a wide range of geometric tasks.
1 code implementation • ICCV 2021 • Shuzhe Wang, Zakaria Laskar, Iaroslav Melekhov, Xiaotian Li, Juho Kannala
For several emerging technologies such as augmented reality, autonomous driving and robotics, visual localization is a critical component.
1 code implementation • 22 Jun 2021 • Otto Seiskari, Pekka Rantalankila, Juho Kannala, Jerry Ylilammi, Esa Rahtu, Arno Solin
We present HybVIO, a novel hybrid approach for combining filtering-based visual-inertial odometry (VIO) with optimization-based SLAM.
2 code implementations • 15 Jun 2021 • Rinu Boney, Alexander Ilin, Juho Kannala
In many control problems that include vision, optimal controls can be inferred from the location of the objects in the scene.
no code implementations • 7 Apr 2021 • Soumya Tripathy, Juho Kannala, Esa Rahtu
Image reenactment is a task where the target object in the source image imitates the motion represented in the driving image.
1 code implementation • 5 Jan 2021 • Yuxin Hou, Arno Solin, Juho Kannala
Flow predictions enable the target view to re-use pixels directly, but can easily lead to distorted results.
1 code implementation • 22 Dec 2020 • Rinu Boney, Alexander Ilin, Juho Kannala, Jarno Seppänen
We experimentally show that planning with naive Monte Carlo tree search does not perform very well in large combinatorial action spaces.
no code implementations • 9 Nov 2020 • Soumya Tripathy, Juho Kannala, Esa Rahtu
However, if the identity differs, the driving facial structures leak to the output distorting the reenactment result.
1 code implementation • 5 Nov 2020 • Rinu Boney, Jussi Sainio, Mikko Kaivola, Arno Solin, Juho Kannala
We validate the platform with reinforcement learning experiments and provide baseline results on a set of benchmark tasks.
1 code implementation • 18 Oct 2020 • Yuxin Hou, Muhammad Kamran Janjua, Juho Kannala, Arno Solin
We propose a method for fusing stereo disparity estimation with movement-induced prior information.
no code implementations • 1 Oct 2020 • Luca Ferranti, Xiaotian Li, Jani Boutellier, Juho Kannala
Camera pose estimation in large-scale environments is still an open question and, despite recent promising results, it may still fail in some situations.
no code implementations • 16 Aug 2020 • Iaroslav Melekhov, Gabriel J. Brostow, Juho Kannala, Daniyar Turmukhambetov
Local features that are robust to both viewpoint and appearance changes are crucial for many computer vision tasks.
no code implementations • 10 Jul 2020 • Zakaria Laskar, Juho Kannala
In low training sample settings, our approach outperforms the fully supervised approach on two challenging image retrieval datasets, ROxford5k and RParis6k \cite{Roxf} with the least possible teacher supervision.
1 code implementation • CVPR 2021 • Jisoo Jeong, Vikas Verma, Minsung Hyun, Juho Kannala, Nojun Kwak
Despite the data labeling cost for the object detection tasks being substantially more than that of the classification tasks, semi-supervised learning methods for object detection have not been studied much.
no code implementations • 20 May 2020 • Luca Ferranti, Kalle Åström, Magnus Oskarsson, Jani Boutellier, Juho Kannala
Given a network of receivers and transmitters, the process of determining their positions from measured pseudoranges is known as network self-calibration.
2 code implementations • NeurIPS 2020 • Ari Heljakka, Yuxin Hou, Juho Kannala, Arno Solin
These networks can faithfully reproduce individual real-world input images like regular autoencoders, but also generate a fused sample from an arbitrary combination of several such images, allowing instantaneous 'style-mixing' and other new applications.
no code implementations • 12 Oct 2019 • Rinu Boney, Juho Kannala, Alexander Ilin
Model-based reinforcement learning could enable sample-efficient learning by quickly acquiring rich knowledge about the world and using it to improve behaviour without additional data.
no code implementations • 25 Sep 2019 • Vikas Verma, Meng Qu, Alex Lamb, Yoshua Bengio, Juho Kannala, Jian Tang
We present GraphMix, a regularization technique for Graph Neural Network based semi-supervised object classification, leveraging the recent advances in the regularization of classical deep neural networks.
1 code implementation • 25 Sep 2019 • Vikas Verma, Meng Qu, Kenji Kawaguchi, Alex Lamb, Yoshua Bengio, Juho Kannala, Jian Tang
We present GraphMix, a regularization method for Graph Neural Network based semi-supervised object classification, whereby we propose to train a fully-connected network jointly with the graph neural network via parameter sharing and interpolation-based regularization.
Ranked #1 on Node Classification on Pubmed random partition
no code implementations • CVPR 2020 • Xiaotian Li, Shuzhe Wang, Yi Zhao, Jakob Verbeek, Juho Kannala
In this work, we present a new hierarchical scene coordinate network to predict pixel scene coordinates in a coarse-to-fine manner from a single RGB image.
3 code implementations • 16 Jun 2019 • Alex Lamb, Vikas Verma, Kenji Kawaguchi, Alexander Matyasko, Savya Khosla, Juho Kannala, Yoshua Bengio
Adversarial robustness has become a central goal in deep learning, both in the theory and the practice.
no code implementations • 2 Jun 2019 • Santiago Cortés Reina, Yuxin Hou, Juho Kannala, Arno Solin
Modern smartphones have all the sensing capabilities required for accurate and robust navigation and tracking.
2 code implementations • 25 May 2019 • Hamed R. -Tavakoli, Ali Borji, Esa Rahtu, Juho Kannala
Our results suggest that (1) audio is a strong contributing cue for saliency prediction, (2) salient visible sound-source is the natural cause of the superiority of our Audio-Visual model, (3) richer feature representations for the input space leads to more powerful predictions even in absence of more sophisticated saliency decoders, and (4) Audio-Visual model improves over 53. 54\% of the frames predicted by the best Visual model (our baseline).
no code implementations • 15 Apr 2019 • Zakaria Laskar, Iaroslav Melekhov, Hamed R. -Tavakoli, Juha Ylioinas, Juho Kannala
The main contribution is a geometric correspondence verification approach for re-ranking a shortlist of retrieved database images based on their dense pair-wise matching with the query image at a pixel level.
no code implementations • 12 Apr 2019 • Hamed R. -Tavakoli, Esa Rahtu, Juho Kannala, Ali Borji
Extensive experiments over multiple datasets reveal that (1) spatial biases are strong in egocentric videos, (2) bottom-up saliency models perform poorly in predicting gaze and underperform spatial biases, (3) deep features perform better compared to traditional features, (4) as opposed to hand regions, the manipulation point is a strong influential cue for gaze prediction, (5) combining the proposed recurrent model with bottom-up cues, vanishing points and, in particular, manipulation point results in the best gaze prediction accuracy over egocentric videos, (6) the knowledge transfer works best for cases where the tasks or sequences are similar, and (7) task and activity recognition can benefit from gaze prediction.
1 code implementation • ICCV 2019 • Yuxin Hou, Juho Kannala, Arno Solin
The flexibility of the Gaussian process (GP) prior provides adapting memory for fusing information from previous views.
1 code implementation • 12 Apr 2019 • Ari Heljakka, Arno Solin, Juho Kannala
retaining the identity of a face), sharp generated/reconstructed samples in high resolutions, and a well-structured latent space that supports semantic manipulation of the inputs.
1 code implementation • 3 Apr 2019 • Ahti Kalervo, Juha Ylioinas, Markus Häikiö, Antti Karhu, Juho Kannala
Better understanding and modelling of building interiors and the emergence of more impressive AR/VR technology has brought up the need for automatic parsing of floorplan images.
1 code implementation • 3 Apr 2019 • Soumya Tripathy, Juho Kannala, Esa Rahtu
This paper presents a generic face animator that is able to control the pose and expressions of a given face image.
no code implementations • NeurIPS 2019 • Rinu Boney, Norman Di Palo, Mathias Berglund, Alexander Ilin, Juho Kannala, Antti Rasmus, Harri Valpola
Trajectory optimization using a learned model of the environment is one of the core elements of model-based reinforcement learning.
4 code implementations • 9 Mar 2019 • Vikas Verma, Kenji Kawaguchi, Alex Lamb, Juho Kannala, Arno Solin, Yoshua Bengio, David Lopez-Paz
We introduce Interpolation Consistency Training (ICT), a simple and computation efficient algorithm for training Deep Neural Networks in the semi-supervised learning paradigm.
1 code implementation • 6 Feb 2019 • Yuxin Hou, Arno Solin, Juho Kannala
This paper presents a novel method, MaskMVS, to solve depth estimation for unstructured multi-view image-pose pairs.
1 code implementation • 29 Jan 2019 • Aarno Oskar Vuola, Saad Ullah Akram, Juho Kannala
Nuclei segmentation is both an important and in some ways ideal task for modern computer vision methods, e. g. convolutional neural networks.
no code implementations • 24 Jan 2019 • Zakaria Laskar, Hamed R. -Tavakoli, Juho Kannala
The problem is posed as finding the geometric transformation that aligns a given image pair.
no code implementations • 24 Jan 2019 • Zakaria Laskar, Juho Kannala
Convolutional neural networks (CNNs) have been successfully applied to solve the problem of correspondence estimation between semantically related images.
no code implementations • 23 Nov 2018 • Janne Mustaniemi, Juho Kannala, Jiri Matas, Simo Särkkä, Janne Heikkilä
The paper addresses the problem of acquiring high-quality photographs with handheld smartphone cameras in low-light imaging conditions.
4 code implementations • 19 Oct 2018 • Iaroslav Melekhov, Aleksei Tiulpin, Torsten Sattler, Marc Pollefeys, Esa Rahtu, Juho Kannala
This paper addresses the challenge of dense pixel correspondence estimation between two images.
Ranked #2 on Dense Pixel Correspondence Estimation on HPatches
Dense Pixel Correspondence Estimation Optical Flow Estimation +1
1 code implementation • 1 Oct 2018 • Janne Mustaniemi, Juho Kannala, Simo Särkkä, Jiri Matas, Janne Heikkilä
We propose a deblurring method that incorporates gyroscope measurements into a convolutional neural network (CNN).
no code implementations • 15 Aug 2018 • Xiaotian Li, Juha Ylioinas, Jakob Verbeek, Juho Kannala
Image-based camera relocalization is an important problem in computer vision and robotics.
1 code implementation • 10 Aug 2018 • Santiago Cortés, Arno Solin, Juho Kannala
Strapdown inertial navigation systems are sensitive to the quality of the data provided by the accelerometer and gyroscope.
no code implementations • 31 Jul 2018 • Saad Ullah Akram, Talha Qaiser, Simon Graham, Juho Kannala, Janne Heikkilä, Nasir Rajpoot
In this paper, we present a semi-supervised mitosis detection method which is designed to leverage a large number of unlabeled breast cancer WSIs.
1 code implementation • ECCV 2018 • Santiago Cortés, Arno Solin, Esa Rahtu, Juho Kannala
The lack of realistic and open benchmarking datasets for pedestrian visual-inertial odometry has made it hard to pinpoint differences in published methods.
2 code implementations • 20 Jul 2018 • Luiza Sayfullina, Eric Malmi, Juho Kannala
The disambiguation is formulated as a binary text classification problem where the prediction is made for the potential soft skill based on the context where it occurs.
1 code implementation • 9 Jul 2018 • Ari Heljakka, Arno Solin, Juho Kannala
Instead, we propose the Progressively Growing Generative Autoencoder (PIONEER) network which achieves high-quality reconstruction with $128{\times}128$ images without requiring a GAN discriminator.
1 code implementation • 31 May 2018 • Santiago Cortés Reina, Arno Solin, Juho Kannala
This application paper proposes a model for estimating the parameters on the fly by fusing gyroscope and camera data, both readily available in modern day smartphones.
no code implementations • 22 May 2018 • Janne Mustaniemi, Juho Kannala, Simo Särkkä, Jiri Matas, Janne Heikkilä
It is well-known that motion blur decreases the performance of traditional feature detectors and descriptors.
1 code implementation • 8 May 2018 • Soumya Tripathy, Juho Kannala, Esa Rahtu
In this paper, we propose a new general purpose image-to-image translation model that is able to utilize both paired and unpaired training data simultaneously.
no code implementations • 24 Apr 2018 • Markus Ylimäki, Juho Kannala, Janne Heikkilä
Then, the original depth maps are re-registered to the fused point cloud to refine the original camera extrinsic parameters.
no code implementations • 20 Feb 2018 • Yao Lu, Jack Valmadre, Heng Wang, Juho Kannala, Mehrtash Harandi, Philip H. S. Torr
State-of-the-art neural network models estimate large displacement optical flow in multi-resolution and use warping to propagate the estimation between two resolutions.
2 code implementations • 14 Feb 2018 • Ari Heljakka, Arno Solin, Juho Kannala
By treating the age phases as a sequence of image domains, we construct a chain of transformers that map images from one age domain to the next.
no code implementations • 9 Feb 2018 • Xiaotian Li, Juha Ylioinas, Juho Kannala
In this paper, instead of in a patch-based manner, we propose to perform the scene coordinate regression in a full-frame manner to make the computation efficient at test time and, more importantly, to add more global context to the regression process to improve the robustness.
no code implementations • 31 Oct 2017 • Iaroslav Melekhov, Juho Kannala, Esa Rahtu
In this work we propose a neural network based image descriptor suitable for image patch matching, which is an important task in many computer vision applications.
no code implementations • 2 Aug 2017 • Arno Solin, Santiago Cortes, Esa Rahtu, Juho Kannala
This paper presents a novel method for visual-inertial odometry.
no code implementations • 31 Jul 2017 • Zakaria Laskar, Iaroslav Melekhov, Surya Kalia, Juho Kannala
The camera location for the query image is obtained via triangulation from two relative translation estimates using a RANSAC based approach.
no code implementations • 16 May 2017 • Yao Lu, Zhirong Yang, Juho Kannala, Samuel Kaski
A key to the problem is learning a representation of relations.
1 code implementation • 9 May 2017 • Saad Ullah Akram, Juho Kannala, Lauri Eklund, Janne Heikkilä
Microscopy imaging plays a vital role in understanding many biological processes in development and disease.
no code implementations • 23 Mar 2017 • Iaroslav Melekhov, Juha Ylioinas, Juho Kannala, Esa Rahtu
In this paper, we propose an encoder-decoder convolutional neural network (CNN) architecture for estimating camera pose (orientation and location) from a single RGB-image.
1 code implementation • 3 Mar 2017 • Zakaria Laskar, Juho Kannala
Particularly, we show that by making the CNN pay attention on the ROI while extracting query image representation leads to significant improvement over the baseline methods on challenging Oxford5k and Paris6k datasets.
1 code implementation • 1 Mar 2017 • Arno Solin, Santiago Cortes, Esa Rahtu, Juho Kannala
Building a complete inertial navigation system using the limited quality data provided by current smartphones has been regarded challenging, if not impossible.
1 code implementation • 5 Feb 2017 • Iaroslav Melekhov, Juha Ylioinas, Juho Kannala, Esa Rahtu
This paper presents a convolutional neural network based approach for estimating the relative pose between two cameras.
1 code implementation • 29 Nov 2016 • Janne Mustaniemi, Juho Kannala, Simo Särkkä, Jiri Matas, Janne Heikkilä
In the process, we also perform a temporal and spatial alignment of the camera and the IMU.
no code implementations • 23 Sep 2016 • Marko Linna, Juho Kannala, Esa Rahtu
In this paper, we present a method for real-time multi-person human pose estimation from video by utilizing convolutional neural networks.
no code implementations • CVPR 2014 • Pekka Rantalankila, Juho Kannala, Esa Rahtu
The parameters of the graph cut problems are learnt in such a manner that they provide complementary sets of regions.
no code implementations • CVPR 2014 • Andrea Vedaldi, Siddharth Mahendran, Stavros Tsogkas, Subhransu Maji, Ross Girshick, Juho Kannala, Esa Rahtu, Iasonas Kokkinos, Matthew B. Blaschko, David Weiss, Ben Taskar, Karen Simonyan, Naomi Saphra, Sammy Mohamed
We show that the collected data can be used to study the relation between part detection and attribute prediction by diagnosing the performance of classifiers that pool information from different parts of an object.
no code implementations • 21 Jun 2013 • Subhransu Maji, Esa Rahtu, Juho Kannala, Matthew Blaschko, Andrea Vedaldi
This paper introduces FGVC-Aircraft, a new dataset containing 10, 000 images of aircraft spanning 100 aircraft models, organised in a three-level hierarchy.