1 code implementation • 25 Mar 2024 • Damien LaRocque, William Guimont-Martin, David-Alexandre Duclos, Philippe Giguère, François Pomerleau
We show that the combination of two TC datasets yields a latent space that can be interpreted with the properties of the terrains.
Ranked #1 on Time Series Classification on BorealTC (Accuracy (5-fold) metric)
1 code implementation • 4 Jul 2023 • William Guimont-Martin, Jean-Michel Fortin, François Pomerleau, Philippe Giguère
Recent works in object detection in LiDAR point clouds mostly focus on predicting bounding boxes around objects.
1 code implementation • 3 Mar 2023 • Amar Ali-bey, Brahim Chaib-Draa, Philippe Giguère
It refers to the process of identifying a place depicted in a query image using only computer vision.
Ranked #2 on Visual Place Recognition on SPED
1 code implementation • 28 Feb 2023 • Amar Ali-bey, Brahim Chaib-Draa, Philippe Giguère
These proxy representations are thus used to construct a global index that encompasses the similarities between all places in the dataset, allowing for highly informative mini-batch sampling at each training iteration.
Ranked #5 on Visual Place Recognition on Nordland
1 code implementation • 31 Oct 2022 • Vincent Grondin, Jean-Michel Fortin, François Pomerleau, Philippe Giguère
Tree perception is an essential building block toward autonomous forestry operations.
1 code implementation • 19 Oct 2022 • Amar Ali-bey, Brahim Chaib-Draa, Philippe Giguère
This paper aims to investigate representation learning for large scale visual place recognition, which consists of determining the location depicted in a query image by referring to a database of reference images.
Ranked #5 on Visual Place Recognition on Pittsburgh-250k-test
1 code implementation • 8 Oct 2022 • Vincent Grondin, François Pomerleau, Philippe Giguère
In this work, we propose to use simulated forest environments to automatically generate 43 k realistic synthetic images with pixel-level annotations, and use it to train deep learning algorithms for tree detection.
1 code implementation • 3 Mar 2022 • Jean-Michel Fortin, Olivier Gamache, Vincent Grondin, François Pomerleau, Philippe Giguère
Using our dataset, we then compare three neural network architectures on the task of individual logs detection and segmentation; two region-based methods and one attention-based method.
no code implementations • 1 Jan 2021 • Alexandre Lemire Paquin, Brahim Chaib-Draa, Philippe Giguère
We prove new generalization bounds for stochastic gradient descent for both the convex and non-convex case.
no code implementations • 29 Jan 2020 • Patrick Dallaire, Luca Ambrogioni, Ludovic Trottier, Umut Güçlü, Max Hinne, Philippe Giguère, Brahim Chaib-Draa, Marcel van Gerven, Francois Laviolette
This paper introduces the Indian Chefs Process (ICP), a Bayesian nonparametric prior on the joint space of infinite directed acyclic graphs (DAGs) and orders that generalizes Indian Buffet Processes.
1 code implementation • 6 Dec 2019 • Martin Robert, Patrick Dallaire, Philippe Giguère
Oftentimes, it relies on collections of visual signatures based on descriptors, such as SIFT or SURF.
1 code implementation • 26 Nov 2019 • Jean-Philippe Mercier, Mathieu Garon, Philippe Giguère, Jean-François Lalonde
In this context, we propose a generic 2D object instance detection approach that uses example viewpoints of the target object at test time to retrieve its 2D location in RGB images, without requiring any additional training (i. e. fine-tuning) step.
no code implementations • 26 Oct 2019 • Charles-Éric Noël Laflamme, François Pomerleau, Philippe Giguère
This report is a survey of the different autonomous driving datasets which have been published up to date.
no code implementations • 25 Sep 2019 • Alexandre Lemire Paquin, Brahim Chaib-Draa, Philippe Giguère
One approach to try to exploit such understanding would be to then make the bias explicit in the loss function.
1 code implementation • 6 May 2019 • Emanuele Palazzolo, Jens Behley, Philipp Lottes, Philippe Giguère, Cyrill Stachniss
For localization and mapping, we employ an efficient direct tracking on the truncated signed distance function (TSDF) and leverage color information encoded in the TSDF to estimate the pose of the sensor.
Robotics
no code implementations • 6 Mar 2019 • Alexandre Gariépy, Jean-Christophe Ruel, Brahim Chaib-Draa, Philippe Giguère
To this effect, we present Grasp Quality Spatial Transformer Network (GQ-STN), a one-shot grasp detection network.
no code implementations • 2 Oct 2018 • David Landry, François Pomerleau, Philippe Giguère
The fusion of Iterative Closest Point (ICP) reg- istrations in existing state estimation frameworks relies on an accurate estimation of their uncertainty.
Robotics
no code implementations • 2 Oct 2018 • Philippe Babin, Philippe Giguère, François Pomerleau
However, without a large scale comparison of solutions to filter outliers, it is becoming tedious to select an appropriate algorithm for a given application.
Robotics
no code implementations • 18 Jun 2018 • Jean-Philippe Mercier, Chaitanya Mitash, Philippe Giguère, Abdeslam Boularias
We then show that the performance of the detector can be substantially improved by using a small set of weakly annotated real images, where a human provides only a list of objects present in each image without indicating the location of the objects.
2 code implementations • 2 Mar 2018 • Mathieu Carpentier, Philippe Giguère, Jonathan Gaudreault
Tree species identification using bark images is a challenging problem that could prove useful for many forestry related tasks.
1 code implementation • ICLR 2018 • Ludovic Trottier, Philippe Giguère, Brahim Chaib-Draa
We show that CNNs connected with our Deep Collaboration obtain better accuracy on facial landmark detection with related tasks.
no code implementations • 2 Jun 2016 • Ludovic Trottier, Philippe Giguère, Brahim Chaib-Draa
The ability to grasp ordinary and potentially never-seen objects is an important feature in both domestic and industrial robotics.
no code implementations • 30 May 2016 • Ludovic Trottier, Philippe Giguère, Brahim Chaib-Draa
Object recognition is an important task for improving the ability of visual systems to perform complex scene understanding.
no code implementations • 19 Mar 2015 • Lucas Rioux-Maldague, Philippe Giguère
We applied our technique to American Sign Language fingerspelling classification using a Deep Belief Network, for which our feature extraction technique is tailored.