Filters








8,731 Hits in 3.9 sec

Deep Mixture Density Network for Probabilistic Object Detection [article]

Yihui He, Jianren Wang
2020 arXiv   pre-print
The bounding box borders of an occluded object can have multiple plausible configurations. We propose a deep multivariate mixture of Gaussians model for probabilistic object detection.  ...  Mistakes/uncertainties in object detection could lead to catastrophes when deploying robots in the real world.  ...  RELATED WORK a) Object Detection: Deep convolutional neural networks were first introduced to object detection in R-CNN [10] and Fast R-CNN [11] .  ... 
arXiv:1911.10614v2 fatcat:hm5m4iko3vdu3atky55plqpl2m

Deep Directional Statistics: Pose Estimation with Uncertainty Quantification [article]

Sergey Prokudin, Peter Gehler, Sebastian Nowozin
2018 arXiv   pre-print
In this paper, we propose a novel probabilistic deep learning model for the task of angular regression. Our model uses von Mises distributions to predict a distribution over object pose angle.  ...  Probabilistic deep learning models combine the expressive power of deep learning with uncertainty quantification.  ...  Moreover, instead of assuming a fixed form for the predictive density we allow for very flexible distributions, specified by a deep neural network.  ... 
arXiv:1805.03430v1 fatcat:rfnvorahpbg2pjia3pq4poyh44

Deep Directional Statistics: Pose Estimation with Uncertainty Quantification [chapter]

Sergey Prokudin, Peter Gehler, Sebastian Nowozin
2018 Lecture Notes in Computer Science  
In this paper we propose a novel probabilistic deep learning model for the task of angular regression. Our model uses von Mises distributions to predict a distribution over object pose angle.  ...  Probabilistic deep learning models combine the expressive power of deep learning with uncertainty quantification.  ...  Moreover, instead of assuming a fixed form for the predictive density we allow for flexible multimodal distributions, specified by a deep neural network.  ... 
doi:10.1007/978-3-030-01240-3_33 fatcat:kqmpurcxm5fjjb6qx7wck4a4ge

Vision-as-Inverse-Graphics: Obtaining a Rich 3D Explanation of a Scene from a Single Image

Lukasz Romaszko, Christopher K. I. Williams, Pol Moreno, Pushmeet Kohli
2017 2017 IEEE International Conference on Computer Vision Workshops (ICCVW)  
For the prediction of the camera latent variables we introduce a novel architecture termed Probabilistic HoughNets (PHNs), which provides a principled approach to combining information from multiple detections  ...  The framework's stages include object detection, the prediction of the camera and lighting variables, and prediction of object-specific variables (shape, appearance and pose).  ...  elements which represent the predictions using a mixture of Gaussians (not bins), and make these predictions with a deep neural network.  ... 
doi:10.1109/iccvw.2017.115 dblp:conf/iccvw/RomaszkoWMK17 fatcat:vfpdg4l4dzgfrmjfyztbfsjsxa

A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving [article]

Di Feng, Ali Harakeh, Steven Waslander, Klaus Dietmayer
2020 arXiv   pre-print
In recent years, deep learning has become the de-facto approach for object detection, and many probabilistic object detectors have been proposed.  ...  First, we provide an overview of generic uncertainty estimation in deep learning, and then systematically survey existing methods and evaluation metrics for probabilistic object detection.  ...  [79] use a deep neural network to directly predict the conditional target density of an energy-based model.  ... 
arXiv:2011.10671v1 fatcat:a7exswrvjfczpln7xojouzklby

Active Learning for Deep Object Detection via Probabilistic Modeling [article]

Jiwoong Choi, Ismail Elezi, Hyuk-Jae Lee, Clement Farabet, Jose M. Alvarez
2021 arXiv   pre-print
Our approach relies on mixture density networks that estimate a probabilistic distribution for each localization and classification head's output.  ...  In this paper, we propose a novel deep active learning approach for object detection.  ...  Mixture density networks have been recently used for several deep learning tasks. The approach of [8] focuses on the regression task for the steering angle.  ... 
arXiv:2103.16130v2 fatcat:2vviqtmpifc53bvh5u57ir724e

Energy-Based Models for Deep Probabilistic Regression [article]

Fredrik K. Gustafsson, Martin Danelljan, Goutam Bhat, Thomas B. Schön
2020 arXiv   pre-print
In our proposed approach, we create an energy-based model of the conditional target density p(y|x), using a deep neural network to predict the un-normalized density from (x,y).  ...  Notably, our model achieves a 2.2% AP improvement over Faster-RCNN for object detection on the COCO dataset, and sets a new state-of-the-art on visual tracking when applied for bounding box estimation.  ...  To allow for multimodal models p(y; φ θ (x)), mixture density networks (MDNs) [3] have also been applied [40, 61] .  ... 
arXiv:1909.12297v4 fatcat:bjn2czgkfjgchfhnxfsubyfoei

Abnormal Human Behavior Detection in Videos: A Review

Huiyu Mu, Ruizhi Sun, Gang Yuan, Yun Wang
2021 Information Technology and Control  
More researches will promotethe development of abnormal human behavior detection, e.g. deep generative network, weakly-supervised.  ...  Modeling human behavior patterns for detecting the abnormal event has become an important domain in recentyears.  ...  Figure 5 The movement inconsistency of the same object in different locations It is noted that probabilistic methods approximated the probability density of the normal samples and detected whether a new  ... 
doi:10.5755/j01.itc.50.3.27864 fatcat:5l6f4gfhazautop6xg46q6v3lm

Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction [article]

Ajay Jain, Sergio Casas, Renjie Liao, Yuwen Xiong, Song Feng, Sean Segal, Raquel Urtasun
2019 arXiv   pre-print
In this work, we propose the discrete residual flow network (DRF-Net), a convolutional neural network for human motion prediction that captures the uncertainty inherent in long-range motion forecasting  ...  Self-driving vehicles plan around both static and dynamic objects, applying predictive models of behavior to estimate future locations of the objects in the environment.  ...  Acknowledgments We would like to thank Abbas Sadat for useful discussions during the development of this research.  ... 
arXiv:1910.08041v1 fatcat:mitmm4irprhjjbm656ed5gdhju

Inference Compilation and Universal Probabilistic Programming [article]

Tuan Anh Le, Atilim Gunes Baydin, Frank Wood
2017 arXiv   pre-print
We introduce a method for using deep neural networks to amortize the cost of inference in models from the family induced by universal probabilistic programming languages, establishing a framework that  ...  combines the strengths of probabilistic programming and deep learning methods.  ...  Acknowledgements We would like to thank Hakan Bilen for his help with the MatConvNet setup and showing us how to use his Fast R-CNN implementation and Tom Rainforth for his helpful advice.  ... 
arXiv:1610.09900v2 fatcat:e6cyizuc2fcypltpcr3x6pucoa

A deep mixture density network for outlier-corrected interpolation of crowd-sourced weather data [article]

Charlie Kirkwood, Theo Economou, Henry Odbert, Nicolas Pugeault
2022 arXiv   pre-print
Our approach implements a Gaussian-uniform mixture density network whose dual purposes - modelling the phenomenon of interest, and learning to classify and ignore outliers - are achieved simultaneously  ...  To this end, here we present a Bayesian deep learning approach for spatio-temporal modelling of environmental variables with automatic outlier detection.  ...  Our deep mixture density network approach to outlier classification unifies outlier detection and correction as part of a same single probabilistic data modelling process, which provides a more streamlined  ... 
arXiv:2201.10544v1 fatcat:sxxq7ph3zbaddmt7j2dxtkxldm

Photometric redshift estimation via deep learning

A. D'Isanto, K. L. Polsterer
2018 Astronomy and Astrophysics  
A modified version of a deep convolutional network was combined with a mixture density network.  ...  We have adopted a feature based random forest and a plain mixture density network to compare performances on experiments with data from SDSS (DR9).  ...  We would like to thank Nikos Gianniotis and Erica Hopkins for proofreading and commenting on this work.  ... 
doi:10.1051/0004-6361/201731326 fatcat:7nlodhpn6balhli6hp63f72ieu

Predictive Uncertainty Quantification with Compound Density Networks [article]

Agustinus Kristiadi, Sina Däubener, Asja Fischer
2019 arXiv   pre-print
Despite the huge success of deep neural networks (NNs), finding good mechanisms for quantifying their prediction uncertainty is still an open problem.  ...  The resulting class of models can be seen as the continuous counterpart to mixture density networks and is therefore referred to as compound density networks (CDNs).  ...  Mixture Density Networks Let D = {x n , y n } N n=1 be an i.i.d dataset.  ... 
arXiv:1902.01080v2 fatcat:62fg3xgdvrfuxctirgmcxqv254

Object Detection as Probabilistic Set Prediction [article]

Georg Hess, Christoffer Petersson, Lennart Svensson
2022 arXiv   pre-print
Accurate uncertainty estimates are essential for deploying deep object detectors in safety-critical systems.  ...  Using the negative log-likelihood for random finite sets, we present a proper scoring rule for evaluating and training probabilistic object detectors.  ...  Computational resources were provided by the Swedish National Infrastructure for Computing at C3SE and NSC, partially funded by the Swedish Research Council, grant agreement no. 2018-05973.  ... 
arXiv:2203.07980v3 fatcat:gevurxjb6rbrxnzodwpvlooil4

A Tetrahedron-Based Heat Flux Signature for Cortical Thickness Morphometry Analysis [chapter]

Yonghui Fan, Gang Wang, Natasha Lepore, Yalin Wang
2018 Lecture Notes in Computer Science  
for Fast Probabilistic Diffeomorphic Registration 286 Conditional Entropy as a Supervised Primitive Segmentation Loss Function 290 Adversarial Similarity Network for Evaluating Image Alignment in Deep  ...  314 Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound 316 SPNet: Shape Prediction using a Fully Convolutional Neural Network 317 Modeling Longitudinal  ... 
doi:10.1007/978-3-030-00931-1_48 pmid:30338317 pmcid:PMC6191198 fatcat:dqhvpm5xzrdqhglrfftig3qejq
« Previous Showing results 1 — 15 out of 8,731 results