Monte Carlo Dropout Ensembles for Robust Illumination Estimation

• URL: http://arxiv.org/abs/2007.10114v1
• Date: Mon, 20 Jul 2020 13:56:14 GMT
• Title: Monte Carlo Dropout Ensembles for Robust Illumination Estimation
• Authors: Firas Laakom, Jenni Raitoharju, Alexandros Iosifidis, Jarno Nikkanen and Moncef Gabbouj
• Abstract summary: Computational color constancy is a preprocessing step used in many camera systems. We propose to aggregate different deep learning methods according to their output uncertainty. The proposed framework leads to state-of-the-art performance on INTEL-TAU dataset.
• Score: 94.14796147340041
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Computational color constancy is a preprocessing step used in many camera systems. The main aim is to discount the effect of the illumination on the colors in the scene and restore the original colors of the objects. Recently, several deep learning-based approaches have been proposed to solve this problem and they often led to state-of-the-art performance in terms of average errors. However, for extreme samples, these methods fail and lead to high errors. In this paper, we address this limitation by proposing to aggregate different deep learning methods according to their output uncertainty. We estimate the relative uncertainty of each approach using Monte Carlo dropout and the final illumination estimate is obtained as the sum of the different model estimates weighted by the log-inverse of their corresponding uncertainties. The proposed framework leads to state-of-the-art performance on INTEL-TAU dataset.

Related papers

• EnsIR: An Ensemble Algorithm for Image Restoration via Gaussian Mixture Models [70.60381055741391]
  Image restoration challenges related to illposed problems, resulting in deviations between single model predictions and ground-truths. Ensemble learning aims to address these deviations by combining the predictions of multiple base models. We employ an expectation (EM)-based algorithm to estimate ensemble weights for prediction candidates. Our algorithm is model-agnostic and training-free, allowing seamless integration and enhancement of various pre-trained image restoration models.
  arXiv  Detail & Related papers  (2024-10-30T12:16:35Z)
• Single Image Depth Prediction Made Better: A Multivariate Gaussian Take [163.14849753700682]
  We introduce an approach that performs continuous modeling of per-pixel depth. Our method's accuracy (named MG) is among the top on the KITTI depth-prediction benchmark leaderboard.
  arXiv  Detail & Related papers  (2023-03-31T16:01:03Z)
• Unsupervised Light Field Depth Estimation via Multi-view Feature Matching with Occlusion Prediction [15.421219881815956]
  It is costly to obtain sufficient depth labels for supervised training. In this paper, we propose an unsupervised framework to estimate depth from LF images.
  arXiv  Detail & Related papers  (2023-01-20T06:11:17Z)
• RISP: Rendering-Invariant State Predictor with Differentiable Simulation and Rendering for Cross-Domain Parameter Estimation [110.4255414234771]
  Existing solutions require massive training data or lack generalizability to unknown rendering configurations. We propose a novel approach that marries domain randomization and differentiable rendering gradients to address this problem. Our approach achieves significantly lower reconstruction errors and has better generalizability among unknown rendering configurations.
  arXiv  Detail & Related papers  (2022-05-11T17:59:51Z)
• Query-Efficient Black-box Adversarial Attacks Guided by a Transfer-based Prior [50.393092185611536]
  We consider the black-box adversarial setting, where the adversary needs to craft adversarial examples without access to the gradients of a target model. Previous methods attempted to approximate the true gradient either by using the transfer gradient of a surrogate white-box model or based on the feedback of model queries. We propose two prior-guided random gradient-free (PRGF) algorithms based on biased sampling and gradient averaging.
  arXiv  Detail & Related papers  (2022-03-13T04:06:27Z)
• Bayesian deep learning of affordances from RGB images [5.939410304994348]
  We present a deep learning method to predict the affordances available in the environment directly from RGB images. Based on previous work on socially accepted affordances, our model is based on a multiscale CNN that combines local and global information from the object and the full image. Our results show a marginal better performance of deep ensembles, compared to MC-dropout on the Brier score and the Expected Error.
  arXiv  Detail & Related papers  (2021-09-27T07:39:47Z)
• The Cube++ Illumination Estimation Dataset [50.58610459038332]
  A new illumination estimation dataset is proposed in this paper. It consists of 4890 images with known illumination colors as well as with additional semantic data. The dataset can be used for training and testing of methods that perform single or two-illuminant estimation.
  arXiv  Detail & Related papers  (2020-11-19T18:50:08Z)
• The Shooting Regressor; Randomized Gradient-Based Ensembles [0.0]
  An ensemble method is introduced that utilizes randomization and loss function gradients to compute a prediction. Multiple weakly-correlated estimators approximate the gradient at randomly sampled points on the error surface and are aggregated into a final solution.
  arXiv  Detail & Related papers  (2020-09-14T03:20:59Z)
• Pose Proposal Critic: Robust Pose Refinement by Learning Reprojection Errors [17.918364675642998]
  We focus our attention on pose refinement, and show how to push the state-of-the-art further in the case of partial occlusions. The proposed pose refinement method leverages on a simplified learning task, where a CNN is trained to estimate the reprojection error between an observed and a rendered image. Current state-of-the-art results are outperformed for two out of three metrics on the Occlusion LINEMOD benchmark, while performing on-par for the final metric.
  arXiv  Detail & Related papers  (2020-05-13T11:46:04Z)
• A Multi-Hypothesis Approach to Color Constancy [22.35581217222978]
  Current approaches frame the color constancy problem as learning camera specific illuminant mappings. We propose a Bayesian framework that naturally handles color constancy ambiguity via a multi-hypothesis strategy. Our method provides state-of-the-art accuracy on multiple public datasets while maintaining real-time execution.
  arXiv  Detail & Related papers  (2020-02-28T18:05:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.