Confidence Adaptive Anytime Pixel-Level Recognition
- URL: http://arxiv.org/abs/2104.00749v1
- Date: Thu, 1 Apr 2021 20:01:57 GMT
- Title: Confidence Adaptive Anytime Pixel-Level Recognition
- Authors: Zhuang Liu, Trevor Darrell, Evan Shelhamer
- Abstract summary: Anytime inference requires a model to make a progression of predictions which might be halted at any time.
We propose the first unified and end-to-end model approach for anytime pixel-level recognition.
- Score: 86.75784498879354
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Anytime inference requires a model to make a progression of predictions which
might be halted at any time. Prior research on anytime visual recognition has
mostly focused on image classification. We propose the first unified and
end-to-end model approach for anytime pixel-level recognition. A cascade of
"exits" is attached to the model to make multiple predictions and direct
further computation. We redesign the exits to account for the depth and spatial
resolution of the features for each exit. To reduce total computation, and make
full use of prior predictions, we develop a novel spatially adaptive approach
to avoid further computation on regions where early predictions are already
sufficiently confident. Our full model with redesigned exit architecture and
spatial adaptivity enables anytime inference, achieves the same level of final
accuracy, and even significantly reduces total computation. We evaluate our
approach on semantic segmentation and human pose estimation. On Cityscapes
semantic segmentation and MPII human pose estimation, our approach enables
anytime inference while also reducing the total FLOPs of its base models by
44.4% and 59.1% without sacrificing accuracy. As a new anytime baseline, we
measure the anytime capability of deep equilibrium networks, a recent class of
model that is intrinsically iterative, and we show that the
accuracy-computation curve of our architecture strictly dominates it.
Related papers
- Temporal Generalization Estimation in Evolving Graphs [35.859484484574956]
Graph Neural Networks (GNNs) are widely deployed in vast fields, but they often struggle to maintain accurate representations as graphs evolve.
We theoretically establish a lower bound, proving that under mild conditions, representation distortion inevitably occurs over time.
We introduce Smart, a straightforward and effective baseline enhanced by an adaptive feature extractor through self-supervised graph reconstruction.
arXiv Detail & Related papers (2024-04-07T14:19:22Z) - Towards Anytime Classification in Early-Exit Architectures by Enforcing
Conditional Monotonicity [5.425028186820756]
Anytime algorithms are well-suited to environments in which computational budgets are dynamic.
We show that current early-exit networks are not directly applicable to anytime settings.
We propose an elegant post-hoc modification, based on the Product-of-Experts, that encourages an early-exit network to become gradually confident.
arXiv Detail & Related papers (2023-06-05T07:38:13Z) - 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) - Adaptive Sparse Gaussian Process [0.0]
We propose the first adaptive sparse Gaussian Process (GP) able to address all these issues.
We first reformulate a variational sparse GP algorithm to make it adaptive through a forgetting factor.
We then propose updating a single inducing point of the sparse GP model together with the remaining model parameters every time a new sample arrives.
arXiv Detail & Related papers (2023-02-20T21:34:36Z) - Dynamic Iterative Refinement for Efficient 3D Hand Pose Estimation [87.54604263202941]
We propose a tiny deep neural network of which partial layers are iteratively exploited for refining its previous estimations.
We employ learned gating criteria to decide whether to exit from the weight-sharing loop, allowing per-sample adaptation in our model.
Our method consistently outperforms state-of-the-art 2D/3D hand pose estimation approaches in terms of both accuracy and efficiency for widely used benchmarks.
arXiv Detail & Related papers (2021-11-11T23:31:34Z) - PDC-Net+: Enhanced Probabilistic Dense Correspondence Network [161.76275845530964]
Enhanced Probabilistic Dense Correspondence Network, PDC-Net+, capable of estimating accurate dense correspondences.
We develop an architecture and an enhanced training strategy tailored for robust and generalizable uncertainty prediction.
Our approach obtains state-of-the-art results on multiple challenging geometric matching and optical flow datasets.
arXiv Detail & Related papers (2021-09-28T17:56:41Z) - Probabilistic Modeling for Human Mesh Recovery [73.11532990173441]
This paper focuses on the problem of 3D human reconstruction from 2D evidence.
We recast the problem as learning a mapping from the input to a distribution of plausible 3D poses.
arXiv Detail & Related papers (2021-08-26T17:55:11Z) - Improving Uncertainty Calibration via Prior Augmented Data [56.88185136509654]
Neural networks have proven successful at learning from complex data distributions by acting as universal function approximators.
They are often overconfident in their predictions, which leads to inaccurate and miscalibrated probabilistic predictions.
We propose a solution by seeking out regions of feature space where the model is unjustifiably overconfident, and conditionally raising the entropy of those predictions towards that of the prior distribution of the labels.
arXiv Detail & Related papers (2021-02-22T07:02:37Z) - RecSal : Deep Recursive Supervision for Visual Saliency Prediction [2.223733768286313]
Saliency prediction datasets can be used to create more information for each stimulus than just a final aggregate saliency map.
We show that our best method outperforms previous state-of-the-art methods with 50-80% fewer parameters.
arXiv Detail & Related papers (2020-08-31T17:08:34Z)
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.