Filters








6,696 Hits in 6.1 sec

End-to-End Semi-Supervised Object Detection with Soft Teacher [article]

Mengde Xu, Zheng Zhang, Han Hu, Jianfeng Wang, Lijuan Wang, Fangyun Wei, Xiang Bai, Zicheng Liu
2021 arXiv   pre-print
This paper presents an end-to-end semi-supervised object detection approach, in contrast to previous more complex multi-stage methods.  ...  The end-to-end training gradually improves pseudo label qualities during the curriculum, and the more and more accurate pseudo labels in turn benefit object detection training.  ...  Methodology End-to-End Pseudo-Labeling Framework We first introduce the end-to-end framework for pseudolabel based semi-supervised object detection.  ... 
arXiv:2106.09018v3 fatcat:evjbqibhqrhd3m6fipbxjeg4n4

Toward Semi-Supervised Graphical Object Detection in Document Images

Goutham Kallempudi, Khurram Azeem Hashmi, Alain Pagani, Marcus Liwicki, Didier Stricker, Muhammad Zeshan Afzal
2022 Future Internet  
This paper presents an end-to-end semi-supervised framework for graphical object detection in scanned document images to address this limitation.  ...  As deep learning techniques for object detection become increasingly successful, many supervised deep neural network-based methods have been introduced to recognize graphical objects in documents.  ...  Hence, the primary objective of this paper is to leverage this novel end-to-end semi-supervised framework on graphical page object detection.  ... 
doi:10.3390/fi14060176 fatcat:bmls5pqww5gorbgpc4s3kjtdwa

Scale-Equivalent Distillation for Semi-Supervised Object Detection [article]

Qiushan Guo, Yao Mu, Jianyu Chen, Tianqi Wang, Yizhou Yu, Ping Luo
2022 arXiv   pre-print
Recent Semi-Supervised Object Detection (SS-OD) methods are mainly based on self-training, i.e., generating hard pseudo-labels by a teacher model on unlabeled data as supervisory signals.  ...  Although they achieved certain success, the limited labeled data in semi-supervised learning scales up the challenges of object detection.  ...  To overcome the challenges motioned above, we propose Scale-Equivalent Distillation (SED), a simple yet effective end-to-end semi-supervised learning framework for object detection.  ... 
arXiv:2203.12244v2 fatcat:otalw2k6bbcgnhtctco65txv6y

Humble Teachers Teach Better Students for Semi-Supervised Object Detection [article]

Yihe Tang, Weifeng Chen, Yijun Luo, Yuting Zhang
2021 arXiv   pre-print
We propose a semi-supervised approach for contemporary object detectors following the teacher-student dual model framework.  ...  Compared to the recent state-of-the-art -- STAC, which uses hard labels on sparsely selected hard pseudo samples, the teacher in our model exposes richer information to the student with soft-labels on  ...  to the overall performance of our semi-supervised object detection method.  ... 
arXiv:2106.10456v1 fatcat:fuzhmre2i5annjfjca6i67smom

Federated Semi-Supervised Domain Adaptation via Knowledge Transfer [article]

Madhureeta Das, Xianhao Chen, Xiaoyong Yuan, Lan Zhang
2022 arXiv   pre-print
This paper proposes an innovative approach to achieve SSDA over multiple distributed and confidential datasets, named by Federated Semi-Supervised Domain Adaptation (FSSDA).  ...  Given the rapidly changing machine learning environments and expensive data labeling, semi-supervised domain adaptation (SSDA) is imperative when the labeled data from the source domain is statistically  ...  snowy day object detection).  ... 
arXiv:2207.10727v2 fatcat:rsd2a5bt7ndxdgutdxgk5t7uiy

PseCo: Pseudo Labeling and Consistency Training for Semi-Supervised Object Detection [article]

Gang Li, Xiang Li, Yujie Wang, Yichao Wu, Ding Liang, Shanshan Zhang
2022 arXiv   pre-print
In this paper, we delve into two key techniques in Semi-Supervised Object Detection (SSOD), namely pseudo labeling and consistency training.  ...  We observe that these two techniques currently neglect some important properties of object detection, hindering efficient learning on unlabeled data.  ...  Semi-supervised learning in object detection.  ... 
arXiv:2203.16317v2 fatcat:abzun6cgv5gt5eibw3ngz5cwca

Mind the Gap: Polishing Pseudo labels for Accurate Semi-supervised Object Detection [article]

Lei Zhang, Yuxuan Sun, Wei Wei
2022 arXiv   pre-print
., categories and bounding boxes) of unannotated objects produced by a teacher detector have underpinned much of recent progress in semi-supervised object detection (SSOD).  ...  Moreover, such a scheme can be seamlessly plugged into the existing SSOD framework for joint end-to-end learning.  ...  Semi-supervised Object Detection Object detection [14, 15, 19, 20, 34] is a fundamental task in computer vision domain.  ... 
arXiv:2207.08185v1 fatcat:vljzfpjrgzhuxhi3krfx3rp64m

DTG-SSOD: Dense Teacher Guidance for Semi-Supervised Object Detection [article]

Gang Li, Xiang Li, Yujie Wang, Yichao Wu, Ding Liang, Shanshan Zhang
2022 arXiv   pre-print
The Mean-Teacher (MT) scheme is widely adopted in semi-supervised object detection (SSOD).  ...  With the proposed INC and RM, we integrate Dense Teacher Guidance into Semi-Supervised Object Detection (termed DTG-SSOD), successfully abandoning sparse pseudo labels and enabling more informative learning  ...  , which inspires us to investigate the optimal supervision form for semi-supervised object detection.  ... 
arXiv:2207.05536v1 fatcat:r3devy5bhjf3tjjwd5mwxvospe

Uncertainty-aware Mean Teacher for Source-free Unsupervised Domain Adaptive 3D Object Detection [article]

Deepti Hegde, Vishwanath Sindagi, Velat Kilic, A. Brinton Cooper, Mark Foster, Vishal Patel
2021 arXiv   pre-print
Effectively, we perform automatic soft-sampling of pseudo-labeled data while aligning predictions from the student and teacher networks.  ...  Leveraging model uncertainty allows the mean teacher network to perform implicit filtering by down-weighing losses corresponding uncertain pseudo-labels.  ...  Mean teacher networks are a popular method for unsupervised, semi-supervised and self-supervised approaches, particularly in object detection.  ... 
arXiv:2109.14651v1 fatcat:5wo4dotrb5ejtnqcgqneagnkyi

Temporal Self-Ensembling Teacher for Semi-Supervised Object Detection [article]

Cong Chen and Shouyang Dong and Ye Tian and Kunlin Cao and Li Liu and Yuanhao Guo
2020 arXiv   pre-print
This paper focuses on Semi-Supervised Object Detection (SSOD). Knowledge Distillation (KD) has been widely used for semi-supervised image classification.  ...  Second, our teacher model ensembles its temporal model weights with the student model weights by an exponential moving average (EMA) which allows the teacher gradually learn from the student.  ...  Semi-supervised object detection A successful trial on semi-supervised object detection using deep learning techniques was the CSD model which adapted the Π model to construct the consistent regularization  ... 
arXiv:2007.06144v3 fatcat:rqvfyha4jracjmnwwgsladwysm

A Selective Survey on Versatile Knowledge Distillation Paradigm for Neural Network Models [article]

Jeong-Hoe Ku, JiHun Oh, YoungYoon Lee, Gaurav Pooniwala, SangJeong Lee
2020 arXiv   pre-print
To this end, we give a brief overview of knowledge distillation and some related works including learning using privileged information(LUPI) and generalized distillation(GD).  ...  This paper aims to provide a selective survey about knowledge distillation(KD) framework for researchers and practitioners to take advantage of it for developing new optimized models in the deep neural  ...  [12] proposed a new end-to-end trainable framework to train compact and fast multi-class object detection networks with improved accuracy using knowledge distillation [6] and hint learning [7] .  ... 
arXiv:2011.14554v1 fatcat:46rruchrlbgfxexsyvujsgprou

Learning Efficient Detector with Semi-supervised Adaptive Distillation [article]

Shitao Tang, Litong Feng, Wenqi Shao, Zhanghui Kuang, Wei Zhang, Yimin Chen
2019 arXiv   pre-print
On the COCO database, semi-supervised adaptive distillation (SAD) makes a student detector with a backbone of ResNet-50 surpasses its teacher with a backbone of ResNet-101, while the student has half of  ...  We propose ADL to address this issue by adaptively mimicking the teacher's logits, with more attention paid on two types of hard samples: hard-to-learn samples predicted by teacher with low certainty and  ...  Figure 1 : Semi-supervised adaptive distillation (SAD) schematic. To begin with, the teacher selects and annotates samples with at least one annotation.  ... 
arXiv:1901.00366v2 fatcat:p2u5twvsgbajbopx6x2j2cmeki

CrossRectify: Leveraging Disagreement for Semi-supervised Object Detection [article]

Chengcheng Ma, Xingjia Pan, Qixiang Ye, Fan Tang, Weiming Dong, Changsheng Xu
2022 arXiv   pre-print
Semi-supervised object detection has recently achieved substantial progress.  ...  In this paper, we propose an effective detection framework named CrossRectify, to obtain accurate pseudo labels by simultaneously training two detectors with different initial parameters.  ...  semi-supervised object detection methods under various data augmentations.  ... 
arXiv:2201.10734v2 fatcat:nai3wods5fhc5jcpryd4gmbmo4

Dynamic Curriculum Learning for Great Ape Detection in the Wild [article]

Xinyu Yang, Tilo Burghardt, Majid Mirmehdi
2022 arXiv   pre-print
In contrast to previous semi-supervised methods, our approach gradually improves detection quality by steering training towards virtuous self-reinforcement.  ...  We propose a novel end-to-end curriculum learning approach that leverages large volumes of unlabelled great ape camera trap footage to improve supervised species detector construction in challenging real-world  ...  In this respect we would also like to thank:  ... 
arXiv:2205.00275v1 fatcat:tne72wjv2jf77pnthv2uirxyyq

Hetero-Modal Learning and Expansive Consistency Constraints for Semi-Supervised Detection from Multi-Sequence Data [article]

Bolin Lai, Yuhsuan Wu, Xiao-Yun Zhou, Peng Wang, Le Lu, Lingyun Huang, Mei Han, Jing Xiao, Heping Hu, Adam P. Harrison
2021 arXiv   pre-print
In this paper, we introduce mean teacher hetero-modal detection (MTHD), which addresses two important gaps in current semi-supervised detection.  ...  Using an anchor-free framework, MTHD formulates a mean teacher approach without such compromises, enforcing consistency on the soft-output of object centers and size.  ...  Semi-Supervised Detection As shown in Fig.1(a) , the proposed semi-supervised detector is based off of the popular mean teacher framework [22] .  ... 
arXiv:2103.12972v1 fatcat:m2i7hx5vdbflhd7zjljoh2hwn4
« Previous Showing results 1 — 15 out of 6,696 results