Corpus ID: 220380967

AutoAssign: Differentiable Label Assignment for Dense Object Detection

@article{Zhu2020AutoAssignDL,
  title={AutoAssign: Differentiable Label Assignment for Dense Object Detection},
  author={Benjin Zhu and Jianfeng Wang and Zhengkai Jiang and Fuhang Zong and Songtao Liu and Zeming Li and Jian Sun},
  journal={ArXiv},
  year={2020},
  volume={abs/2007.03496}
}
In this paper, we propose an anchor-free object detector with a fully differentiable label assignment strategy, named AutoAssign. It automatically determines positive/negative samples by generating positive and negative weight maps to modify each location's prediction dynamically. Specifically, we present a center weighting module to adjust the category-specific prior distributions and a confidence weighting module to adapt the specific assign strategy of each instance. The entire label… Expand
Improving Object Detection by Label Assignment Distillation
TLDR
Colearning LAD is introduced, in which two networks simultaneously learn from scratch and the role of teacher and student are dynamically interchanged, and is more effective than soft-label, but they are complementary. Expand
Probabilistic two-stage detection
TLDR
A probabilistic interpretation of twostage object detection is developed that motivates a number of common empirical training practices and suggests changes to two-stage detection pipelines. Expand
OTA: Optimal Transport Assignment for Object Detection
TLDR
This paper innovatively revisit the label assignment from a global perspective and proposes to formulate the assigning procedure as an Optimal Transport (OT) problem – a well-studied topic in Optimization Theory. Expand
A General Gaussian Heatmap Labeling for Arbitrary-Oriented Object Detection
TLDR
A novel AOOD method called General Gaussian Heatmap Labeling (GGHL) is proposed and an anchor-free object-adaptation label assignment (OLA) strategy is presented to define the positive candidates based on two-dimensional oriented Gaussian heatmaps, which reflect the shape and direction features of arbitrary-oriented objects. Expand
Bootstrap Your Object Detector via Mixed Training
  • Mengde Xu, Zheng Zhang, +5 authors Xiang Bai
  • Computer Science
  • ArXiv
  • 2021
TLDR
This work introduces MixTraining, a new training paradigm for object detection that can improve the performance of existing detectors for free and addresses localization noise and missing labels in human annotations by incorporating pseudo boxes that can compensate for these errors. Expand
ByteTrack: Multi-Object Tracking by Associating Every Detection Box
  • Yifu Zhang, Peize Sun, +5 authors Xinggang Wang
  • Computer Science
  • ArXiv
  • 2021
TLDR
A simple, effective and generic association method, tracking by associating every detection box instead of only the high score ones, for the low score detection boxes, to recover true objects and filter out the background detections. Expand
Efficient DETR: Improving End-to-End Object Detector with Dense Prior
TLDR
This paper investigates that the random initialization of object containers, which include object queries and reference points, is mainly responsible for the requirement of multiple iterations of object detection, and proposes Efficient DETR, a simple and efficient pipeline for end-to-end object detection. Expand
End-to-End Object Detection with Fully Convolutional Network
TLDR
A Prediction-aware One-To-One (POTO) label assignment for classification is introduced to enable end-to-end detection, which obtains comparable performance with NMS and a simple 3D Max Filtering is proposed to utilize the multi-scale features and improve the discriminability of convolutions in the local region. Expand
Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection
TLDR
This paper explores a completely novel and different perspective to perform LQE – based on the learned distributions of the four parameters of the bounding box – and develops a considerably lightweight Distribution-Guided Quality Predictor (DGQP) for reliable LqE based on GFLV1, thus producing GFLv2. Expand
Higher Performance Visual Tracking with Dual-Modal Localization
  • Jinghao Zhou, Bo Li, +5 authors Wanli Ouyang
  • Computer Science
  • ArXiv
  • 2021
TLDR
A dual-modal framework for target localization is proposed, consisting of robust localization suppressing distractors via ONR and the accurate localization attending to the target center precisely via OFC, and a simple but effective score voting strategy to involve adjacent predictions such that the final representation does not commit to a single location. Expand
...
1
2
3
...

References

SHOWING 1-10 OF 35 REFERENCES
Microsoft COCO: Common Objects in Context
We present a new dataset with the goal of advancing the state-of-the-art in object recognition by placing the question of object recognition in the context of the broader question of sceneExpand
Focal Loss for Dense Object Detection
TLDR
This paper proposes to address the extreme foreground-background class imbalance encountered during training of dense detectors by reshaping the standard cross entropy loss such that it down-weights the loss assigned to well-classified examples, and develops a novel Focal Loss, which focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training. Expand
FCOS: Fully Convolutional One-Stage Object Detection
TLDR
For the first time, a much simpler and flexible detection framework achieving improved detection accuracy is demonstrated, and it is hoped that the proposed FCOS framework can serve as a simple and strong alternative for many other instance-level tasks. Expand
FreeAnchor: Learning to Match Anchors for Visual Object Detection
TLDR
A learning-to-match (LTM) method to break IoU restriction, allowing objects to match anchors in a flexible manner, validating the general applicability of learnable object-feature matching mechism for visual object detection. Expand
Bridging the Gap Between Anchor-Based and Anchor-Free Detection via Adaptive Training Sample Selection
TLDR
An Adaptive Training Sample Selection (ATSS) to automatically select positive and negative samples according to statistical characteristics of object significantly improves the performance of anchor-based and anchor-free detectors and bridges the gap between them. Expand
Feature Selective Anchor-Free Module for Single-Shot Object Detection
TLDR
The FSAF module robustly improves the baseline RetinaNet by a large margin under various settings, while introducing nearly free inference overhead, and the resulting best model can achieve a state-of-the-art 44.6% mAP, outperforming all existing single-shot detectors on COCO. Expand
Feature Pyramid Networks for Object Detection
TLDR
This paper exploits the inherent multi-scale, pyramidal hierarchy of deep convolutional networks to construct feature pyramids with marginal extra cost and achieves state-of-the-art single-model results on the COCO detection benchmark without bells and whistles. Expand
You Only Look Once: Unified, Real-Time Object Detection
TLDR
Compared to state-of-the-art detection systems, YOLO makes more localization errors but is less likely to predict false positives on background, and outperforms other detection methods, including DPM and R-CNN, when generalizing from natural images to other domains like artwork. Expand
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
TLDR
This work introduces a Region Proposal Network (RPN) that shares full-image convolutional features with the detection network, thus enabling nearly cost-free region proposals and further merge RPN and Fast R-CNN into a single network by sharing their convolutionAL features. Expand
Learning From Noisy Anchors for One-Stage Object Detection
TLDR
This paper proposes to mitigate noise incurred by imperfect label assignment such that the contributions of anchors are dynamically determined by a carefully constructed cleanliness score associated with each anchor, and conducts extensive experiments on COCO. Expand
...
1
2
3
4
...