A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf
.
Filters
FreeAnchor: Learning to Match Anchors for Visual Object Detection
[article]
2019
arXiv
pre-print
In this study, we propose a learning-to-match approach to break IoU restriction, allowing objects to match anchors in a flexible manner. ...
Modern CNN-based object detectors assign anchors for ground-truth objects under the restriction of object-anchor Intersection-over-Unit (IoU). ...
Conclusion We proposed an elegant and effective approach, referred to as FreeAnchor, for visual object detection. ...
arXiv:1909.02466v2
fatcat:fj2mh5q2ize53kzmohrkrhusuq
iffDetector: Inference-aware Feature Filtering for Object Detection
[article]
2020
arXiv
pre-print
By applying Fourier transform analysis, we demonstrate that the IFF module acts as a negative feedback that theoretically guarantees the stability of feature learning. ...
IFF can be fused with CNN-based object detectors in a plug-and-play manner with negligible computational cost overhead. ...
FreeAnchor [30] leverages spatial inference to enforce feature representation and feature-object matching. ...
arXiv:2006.12708v1
fatcat:sgekt3xikfbd3phv5roa7mat24
Variational Pedestrian Detection
[article]
2021
arXiv
pre-print
assignment procedure in classical object detection methods. ...
Experiments conducted on CrowdHuman and CityPersons datasets show that the proposed algorithm serves as an efficient solution to handle the dense pedestrian detection problem for the case of single-stage ...
Online Anchor Matching The aforementioned detection methods perform the assignment of ground truth to anchors before adjusting the object boxes. ...
arXiv:2104.12389v1
fatcat:2xkpxpp5dnbshkka65xpvf646m
Operationalizing Convolutional Neural Network Architectures for Prohibited Object Detection in X-Ray Imagery
[article]
2021
arXiv
pre-print
Here, we explore the viability of two recent end-to-end object detection CNN architectures, Cascade R-CNN and FreeAnchor, for prohibited item detection by balancing processing time and the impact of image ...
Overall, we achieve maximal detection performance using a FreeAnchor architecture with a ResNet50 backbone, obtaining mean Average Precision (mAP) of 87.7 and 85.8 for using the OPIXray and SIXray benchmark ...
breaks the IoU restriction, which allows objects to match anchors in a flexible manner. ...
arXiv:2110.04906v1
fatcat:dwut5tgi5jdtzgyn6jjtufsrwi
A New Object Detection Method for Object Deviating from Center or Multi Object Crowding
2021
Displays (Guildford)
Firstly, FreeAnchor is introduced on the basis of RetinaNet, which can autonomously learn to match the target category; secondly, ResNeXt50 is taken as the backbone to improve the accuracy without increasing ...
Highlights 1) Proposing an object detection method for object deviating from center or multi object crowding. 2) The functions of PA-FPN, FreeAnchor and Soft-NMS are complementary and the combination of ...
IOU threshold to judge whether choose anchor or not, which can make the network have the ability of autonomous learning anchor and category matching. ...
doi:10.1016/j.displa.2021.102042
fatcat:d7gvbsrqv5fvbcntx74lgnre4q
Multiple Anchor Learning for Visual Object Detection
2020
2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Classification and localization are two pillars of visual object detectors. ...
In an adversarial selection-depression manner, MAL not only pursues optimal solutions but also fully leverages multiple anchors/features to learn a detection model. ...
Conclusion We have proposed an elegant and effective training approach, referred to as Multiple Anchor Learning (MAL), for visual object detection. ...
doi:10.1109/cvpr42600.2020.01022
dblp:conf/cvpr/KeZHYLH20
fatcat:xhsmmkxvlrdsfmwgt63cfw6olu
Multiple Anchor Learning for Visual Object Detection
[article]
2019
arXiv
pre-print
Classification and localization are two pillars of visual object detectors. ...
In an adversarial selection-depression manner, MAL not only pursues optimal solutions but also fully leverages multiple anchors/features to learn a detection model. ...
Conclusion We have proposed an elegant and effective training approach, referred to as Multiple Anchor Learning (MAL), for visual object detection. ...
arXiv:1912.02252v1
fatcat:nwe35ue2nfhmjpbg772weinvtq
Deep Learning Based Electric Pylon Detection in Remote Sensing Images
2020
Remote Sensing
Considering the low efficiency of manual detection, we propose to utilize deep learning methods for electric pylon detection in high-resolution remote sensing images in this paper. ...
The comparative analysis can provide reference for the selection of specific deep learning model in actual electric pylon detection task. ...
The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. ...
doi:10.3390/rs12111857
fatcat:ynqrppvky5gtfcu4regh5kfawm
Learning Balance Feature for Object Detection
2022
Electronics
However, FPN is insufficient for object detection on various scales, especially for small-scale object detection. ...
In the field of studying scale variation, the Feature Pyramid Network (FPN) replaces the image pyramid and has become one of the most popular object detection methods for detecting multi-scale objects. ...
Anchor-free detectors learn to recognize keypoints of instances, such as center points or corner points, rather than using anchor boxes to detect instances. ...
doi:10.3390/electronics11172765
fatcat:tgg6retgn5gnrlt7gknowgxime
AutoAssign: Differentiable Label Assignment for Dense Object Detection
[article]
2020
arXiv
pre-print
Determining positive/negative samples for object detection is known as label assignment. Here we present an anchor-free detector named AutoAssign. ...
To adapt to object appearances, Confidence Weighting is proposed to adjust the specific assign strategy of each instance. ...
FreeAnchor [27] bag of top-k anchor candidates based on IoU for every object and uses a Mean-Max function to weight among selected anchors, and [9] designs another weighting function to eliminate ...
arXiv:2007.03496v3
fatcat:c3j337ele5gtrcjh6ihn7tughy
LLA: Loss-aware Label Assignment for Dense Pedestrian Detection
[article]
2021
arXiv
pre-print
Finally, anchors with top K minimum joint losses for a certain GT box are assigned as its positive anchors. Anchors that are not assigned to any GT box are considered negative. ...
Label assignment has been widely studied in general object detection because of its great impact on detectors' performance. ...
FreeAnchor [35] designs a detection-customized likelihood that takes precision and recall into consideration to tackle the anchor-object matching problem. ...
arXiv:2101.04307v3
fatcat:mq47f4gtkza6xgtjavr4mpnnzq
Probabilistic Anchor Assignment with IoU Prediction for Object Detection
[article]
2020
arXiv
pre-print
In this paper we propose a novel anchor assignment strategy that adaptively separates anchors into positive and negative samples for a ground truth bounding box according to the model's learning status ...
In object detection, determining which anchors to assign as positive or negative samples, known as anchor assignment, has been revealed as a core procedure that can significantly affect a model's performance ...
Figure 7 shows anchor assignment results on a non-COCO image.
Visualization of Detection Results We visualize detection results on COCO minival set in Figure 8 . ...
arXiv:2007.08103v2
fatcat:uuywiwfopbf4nhfn7s66whygtm
OTA: Optimal Transport Assignment for Object Detection
[article]
2021
arXiv
pre-print
Recent advances in label assignment in object detection mainly seek to independently define positive/negative training samples for each ground-truth (gt) object. ...
Concretely, we define the unit transportation cost between each demander (anchor) and supplier (gt) pair as the weighted summation of their classification and regression losses. ...
Classical label assigning strategies commonly adopt predefined rules to match the ground-truth (gt) object or background for each anchor. ...
arXiv:2103.14259v1
fatcat:bnumhi5s5zh3xbspvtusbj4x6i
A Scale-Aware Pyramid Network for Multi-Scale Object Detection in SAR Images
2022
Remote Sensing
Furthermore, a self-learning anchor assignment is set to update hand-crafted anchor assignments to learnable anchor/feature configuration. ...
Over the past few years, CNN-based detectors have advanced sharply in SAR object detection. ...
To solve the problems above, the learning-to-match (LTM) approach [13] is introduced for label assignment in anchor-based detectors to detect SAR objects that exhibit different appearances. ...
doi:10.3390/rs14040973
fatcat:oixxshz37jhktilee7r4alc6aq
Scale Match for Tiny Person Detection
[article]
2019
arXiv
pre-print
Visual object detection has achieved unprecedented ad-vance with the rise of deep convolutional neural networks.However, detecting tiny objects (for example tiny per-sons less than 20 pixels) in large-scale ...
Accordingly, we proposea simple yet effective Scale Match approach to align theobject scales between the two datasets for favorable tiny-object representation. ...
Scale Match is designed as a plug-and-play universal block for object scale processing, which provides a fresh insight for general object detection tasks. ...
arXiv:1912.10664v1
fatcat:yfhi6q3dvvdqdcqhwtffgvjdxi
« Previous
Showing results 1 — 15 out of 60 results