D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement

• URL: http://arxiv.org/abs/2410.13842v1
• Date: Thu, 17 Oct 2024 17:57:01 GMT
• Title: D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement
• Authors: Yansong Peng, Hebei Li, Peixi Wu, Yueyi Zhang, Xiaoyan Sun, Feng Wu,
• Abstract summary: D-FINE is a powerful real-time object detector that achieves outstanding localization precision. D-FINE comprises two key components: Fine-grained Distribution Refinement (FDR) and Global Optimal localization Self-Distillation (GO-LSD) When pretrained on Objects365, D-FINE-L / X attains 57.1% / 59.3% AP, surpassing all existing real-time detectors.
• Score: 37.78880948551719
• License:
• Abstract: We introduce D-FINE, a powerful real-time object detector that achieves outstanding localization precision by redefining the bounding box regression task in DETR models. D-FINE comprises two key components: Fine-grained Distribution Refinement (FDR) and Global Optimal Localization Self-Distillation (GO-LSD). FDR transforms the regression process from predicting fixed coordinates to iteratively refining probability distributions, providing a fine-grained intermediate representation that significantly enhances localization accuracy. GO-LSD is a bidirectional optimization strategy that transfers localization knowledge from refined distributions to shallower layers through self-distillation, while also simplifying the residual prediction tasks for deeper layers. Additionally, D-FINE incorporates lightweight optimizations in computationally intensive modules and operations, achieving a better balance between speed and accuracy. Specifically, D-FINE-L / X achieves 54.0% / 55.8% AP on the COCO dataset at 124 / 78 FPS on an NVIDIA T4 GPU. When pretrained on Objects365, D-FINE-L / X attains 57.1% / 59.3% AP, surpassing all existing real-time detectors. Furthermore, our method significantly enhances the performance of a wide range of DETR models by up to 5.3% AP with negligible extra parameters and training costs. Our code and pretrained models: https://github.com/Peterande/D-FINE.

Related papers

• FA-Depth: Toward Fast and Accurate Self-supervised Monocular Depth Estimation [11.039105169475484]
  Most existing methods often rely on complex models to predict scene depth with high accuracy, resulting in slow inference that is not conducive to deployment. We first designed SmallDepth based on sparsity. Second, to enhance the feature representation ability of SmallDepth during training under the condition of equal complexity during inference, we propose an equivalent transformation module(ETM) Third, to improve the ability of each layer in the case of a fixed SmallDepth to perceive different context information, we propose pyramid loss. Fourth, to further improve the accuracy of SmallDepth, we utilized the proposed function approximation loss (APX) to
  arXiv  Detail & Related papers  (2024-05-17T16:22:52Z)
• Align-DETR: Improving DETR with Simple IoU-aware BCE loss [32.13866392998818]
  We propose a metric, recall of best-regressed samples, to quantitively evaluate the misalignment problem. The proposed loss, IA-BCE, guides the training of DETR to build a strong correlation between classification score and localization precision. To overcome the dramatic decrease in sample quality induced by the sparsity of queries, we introduce a prime sample weighting mechanism.
  arXiv  Detail & Related papers  (2023-04-15T10:24:51Z)
• Q-DETR: An Efficient Low-Bit Quantized Detection Transformer [50.00784028552792]
  We find that the bottlenecks of Q-DETR come from the query information distortion through our empirical analyses. We formulate our DRD as a bi-level optimization problem, which can be derived by generalizing the information bottleneck (IB) principle to the learning of Q-DETR. We introduce a new foreground-aware query matching scheme to effectively transfer the teacher information to distillation-desired features to minimize the conditional information entropy.
  arXiv  Detail & Related papers  (2023-04-01T08:05:14Z)
• Leveraging Unlabeled Data to Predict Out-of-Distribution Performance [63.740181251997306]
  Real-world machine learning deployments are characterized by mismatches between the source (training) and target (test) distributions. In this work, we investigate methods for predicting the target domain accuracy using only labeled source data and unlabeled target data. We propose Average Thresholded Confidence (ATC), a practical method that learns a threshold on the model's confidence, predicting accuracy as the fraction of unlabeled examples.
  arXiv  Detail & Related papers  (2022-01-11T23:01:12Z)
• Recurrent Glimpse-based Decoder for Detection with Transformer [85.64521612986456]
  We introduce a novel REcurrent Glimpse-based decOder (REGO) in this paper. In particular, the REGO employs a multi-stage recurrent processing structure to help the attention of DETR gradually focus on foreground objects. REGO consistently boosts the performance of different DETR detectors by up to 7% relative gain at the same setting of 50 training epochs.
  arXiv  Detail & Related papers  (2021-12-09T00:29:19Z)
• Physics-aware deep neural networks for surrogate modeling of turbulent natural convection [0.0]
  We investigate the use of PINNs surrogate modeling for turbulent Rayleigh-B'enard convection flows. We show how it comes to play as a regularization close to the training boundaries which are zones of poor accuracy for standard PINNs. The predictive accuracy of the surrogate over the entire half a billion DNS coordinates yields errors for all flow variables ranging between [0.3% -- 4%] in the relative L 2 norm.
  arXiv  Detail & Related papers  (2021-03-05T09:48:57Z)
• Inception Convolution with Efficient Dilation Search [121.41030859447487]
  Dilation convolution is a critical mutant of standard convolution neural network to control effective receptive fields and handle large scale variance of objects. We propose a new mutant of dilated convolution, namely inception (dilated) convolution where the convolutions have independent dilation among different axes, channels and layers. We explore a practical method for fitting the complex inception convolution to the data, a simple while effective dilation search algorithm(EDO) based on statistical optimization is developed.
  arXiv  Detail & Related papers  (2020-12-25T14:58:35Z)
• Real-Time Regression with Dividing Local Gaussian Processes [62.01822866877782]
  Local Gaussian processes are a novel, computationally efficient modeling approach based on Gaussian process regression. Due to an iterative, data-driven division of the input space, they achieve a sublinear computational complexity in the total number of training points in practice. A numerical evaluation on real-world data sets shows their advantages over other state-of-the-art methods in terms of accuracy as well as prediction and update speed.
  arXiv  Detail & Related papers  (2020-06-16T18:43:31Z)

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.