2,943 Hits in 7.2 sec

Multiple instance learning on deep features for weakly supervised object detection with extreme domain shifts [article]

Nicolas Gonthier and Saïd Ladjal and Yann Gousseau
2020 arXiv   pre-print
Whereas such task is typically addressed with a domain-specific solution focused on natural images, we show that a simple multiple instance approach applied on pre-trained deep features yields excellent  ...  Weakly supervised object detection (WSOD) using only image-level annotations has attracted a growing attention over the past few years.  ...  In this work, we take interest in weakly supervised object detection in the case of extreme domain shifts, namely non-photographic images, possibly addressing the detection of new, never seen classes.  ... 
arXiv:2008.01178v4 fatcat:eae2lctf6fb3jdm6maprmstvzy

Weakly Supervised Object Localization and Detection: A Survey [article]

Dingwen Zhang, Junwei Han, Gong Cheng, Ming-Hsuan Yang
2021 arXiv   pre-print
In this work, we review (1) classic models, (2) approaches with feature representations from off-the-shelf deep networks, (3) approaches solely based on deep learning, and (4) publicly available datasets  ...  As an emerging and challenging problem in the computer vision community, weakly supervised object localization and detection plays an important role for developing new generation computer vision systems  ...  In addition, the learner may mistakenly label a bounding-box that contains a bicycle as a motorcycle, as these two object categories share many similar features. • Learning with domain shifts: For a certain  ... 
arXiv:2104.07918v1 fatcat:dwl6sjfzibdilnvjnrbifp4uke

Deep Weakly-Supervised Domain Adaptation for Pain Localization in Videos [article]

Gnana Praveen R, Eric Granger, Patrick Cardinal
2020 arXiv   pre-print
In particular, WSDA integrates multiple instance learning into an adversarial deep domain adaptation framework to train an Inflated 3D-CNN (I3D) model such that it can accurately estimate pain intensities  ...  The training process relies on weak target loss, along with domain loss and source loss for domain adaptation of the I3D model.  ...  Multiple Instance Learning (MIL) is one of the widely used approaches for inexact supervision [6] .  ... 
arXiv:1910.08173v2 fatcat:p7huseb36ve37nk4qpmbygnazq

A Comparative Review of Recent Few-Shot Object Detection Algorithms [article]

Leng Jiaxu, Chen Taiyue, Gao Xinbo, Yu Yongtao, Wang Ye, Gao Feng, Wang Yue
2021 arXiv   pre-print
Finally, we conclude with the current status of few-shot object detection, along with potential research directions for this field.  ...  Few-shot object detection, learning to adapt to the novel classes with a few labeled data, is an imperative and long-lasting problem due to the inherent long-tail distribution of real-world data and the  ...  For content completeness, we also present a compendious review of advances in object detection, few-shot learning, semi-supervised learning and weakly-supervised learning.  ... 
arXiv:2111.00201v1 fatcat:preckuym7zarndm4yjc7n4k2oi

Domain Adaptation for Visual Applications: A Comprehensive Survey [article]

Gabriela Csurka
2017 arXiv   pre-print
The aim of this paper is to give an overview of domain adaptation and transfer learning with a specific view on visual applications.  ...  Fourth, we overview the methods that go beyond image categorization, such as object detection or image segmentation, video analyses or learning visual attributes.  ...  Finally, the most promising methods are based on deep learning architectures designed for DA. Shallow methods with deep features.  ... 
arXiv:1702.05374v2 fatcat:5va4oz4evjfhxgxddflpbb6pxi

Recent Advances in Transfer Learning for Cross-Dataset Visual Recognition: A Problem-Oriented Perspective [article]

Jing Zhang and Wanqing Li and Philip Ogunbona and Dong Xu
2019 arXiv   pre-print
This paper takes a problem-oriented perspective and presents a comprehensive review of transfer learning methods, both shallow and deep, for cross-dataset visual recognition.  ...  The comprehensive problem-oriented review of the advances in transfer learning with respect to the problem has not only revealed the challenges in transfer learning for visual recognition, but also the  ...  Along similar lines, there have been efforts focusing on domain shift issues for 2D object recognition applications.  ... 
arXiv:1705.04396v3 fatcat:iknfmppi5zca7ljovdlwvdwluu

Weakly Supervised Learning for Facial Behavior Analysis : A Review [article]

Gnana Praveen R, Eric Granger, Patrick Cardinal
2021 arXiv   pre-print
In this paper, we provide a comprehensive review of weakly supervised learning (WSL) approaches for facial behavior analysis with both categorical as well as dimensional labels along with the challenges  ...  based approaches for many real world applications.However, the performance of deep learning approaches relies on the amount of training data.  ...  WEAKLY SUPERVISED LEARNING FOR FACIAL BEHAVIOR ANALYSIS The category of machine learning approaches which deals with weakly annotated data are termed as "Weakly Supervised Learning (WSL)".  ... 
arXiv:2101.09858v1 fatcat:fv3nwbr43vfvrlf655jipvhui4

Adapting pedestrian detectors to new domains: A comprehensive review

Kyaw Kyaw Htike, David Hogg
2016 Engineering applications of artificial intelligence  
In this paper, we survey the most relevant and important state-of-the-art results for domain adaptation for image and video data, with a particular focus on pedestrian detection.  ...  There is a real need to review and analyse critically the state-of-the-art domain adaptation algorithms, especially in the area of object and pedestrian detection.  ...  Weakly supervised learning for object detection Galleguillos et al. [64] propose a weakly supervised approach to learn object detectors given weakly labelled images.  ... 
doi:10.1016/j.engappai.2016.01.029 fatcat:ec5u3whdv5fexieznfjrpqx6nm

Bridging Gap between Image Pixels and Semantics via Supervision: A Survey [article]

Jiali Duan, C.-C. Jay Kuo
2022 arXiv   pre-print
Experiences are drawn from two application domains to illustrate this point: 1) object detection and 2) metric learning for content-based image retrieval (CBIR).  ...  Then, it summarizes various supervision methods to bridge the semantic gap in the context of object detection and metric learning.  ...  Weakly supervised object localization with multi-fold multiple instance learning. IEEE transactions on pattern analysis and machine intelligence, 39(1):189-203, 2016. 10 [25] Papers With Code.  ... 
arXiv:2107.13757v3 fatcat:dw4c74c3h5bvlmzmxugeh5aela

Multi-Source Domain Adaptation for Object Detection [article]

Xingxu Yao, Sicheng Zhao, Pengfei Xu, Jufeng Yang
2021 arXiv   pre-print
To reduce annotation labor associated with object detection, an increasing number of studies focus on transferring the learned knowledge from a labeled source domain to another unlabeled target domain.  ...  First, we propose a hierarchical feature alignment strategy to conduct strong and weak alignments for low- and high-level features, respectively, considering their different effects for object detection  ...  At the same time, the supervised learning for object detection for each labeled source domain is conducted in the corresponding source subnet. Low-level feature alignment.  ... 
arXiv:2106.15793v1 fatcat:7bq2o3g6t5h73hd4oriwymgfem

Transfer Adaptation Learning: A Decade Survey [article]

Lei Zhang, Xinbo Gao
2020 arXiv   pre-print
Broader solutions of transfer adaptation learning being created by researchers are identified, i.e., instance re-weighting adaptation, feature adaptation, classifier adaptation, deep network adaptation  ...  Conventional machine learning aims to find a model with the minimum expected risk on test data by minimizing the regularized empirical risk on the training data, which, however, supposes that the training  ...  ACKNOWLEDGMENT The author would like to thank the pioneer researchers in transfer learning, domain adaptation and other related fields. The author would also like to thank Dr. Mingsheng Long and Dr.  ... 
arXiv:1903.04687v2 fatcat:wurprqieffalnnp6isfkhh5y5i

Recent Advances in Transfer Learning for Cross-Dataset Visual Recognition

Jing Zhang, Wanqing Li, Philip Ogunbona, Dong Xu
2019 ACM Computing Surveys  
The comprehensive problem-oriented review of the advances in transfer learning with respect to the problem has not only revealed the challenges in transfer learning for visual recognition but also the  ...  The comprehensive problem-oriented review of the advances in transfer learning with respect to the problem has not only revealed the challenges in transfer learning for visual recognition, but also the  ...  Along similar lines, there have been efforts focusing on domain shift issues for 2D object recognition applications.  ... 
doi:10.1145/3291124 fatcat:thjzho3xsnfalprmkquldhwpvm

Deep Visual Domain Adaptation: A Survey [article]

Mei Wang, Weihong Deng
2018 arXiv   pre-print
Compared to conventional methods, which learn shared feature subspaces or reuse important source instances with shallow representations, deep domain adaption methods leverage deep networks to learn more  ...  In this paper, we provide a comprehensive survey of deep domain adaptation methods for computer vision applications with four major contributions.  ...  multiple instance loss.  ... 
arXiv:1802.03601v4 fatcat:d5hwwecipjfjzmh7725lmepzfe

WELDON: Weakly Supervised Learning of Deep Convolutional Neural Networks

Thibaut Durand, Nicolas Thome, Matthieu Cord
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
In this paper, we introduce a novel framework for WEakly supervised Learning of Deep cOnvolutional neural Networks (WELDON).  ...  Firstly, WELDON leverages recent improvements on the Multiple Instance Learning paradigm, i.e. negative evidence scoring and top instance selection.  ...  Our method exploits to the full extend deep CNN strategy in multiple instance learning framework to efficiently deal with weak supervision.  ... 
doi:10.1109/cvpr.2016.513 dblp:conf/cvpr/DurandTC16 fatcat:2mqnc6l7jzhmhgjr55vispq4he

Weakly Supervised Domain Adaptation for Built-up Region Segmentation in Aerial and Satellite Imagery [article]

Javed Iqbal, Mohsen Ali
2020 arXiv   pre-print
We thoroughly study the limitations of existing domain adaptation methods and propose a weakly-supervised adaptation strategy where we assume image-level labels are available for the target domain.  ...  The diverse nature of aerial and satellite imagery and lack of labeled data covering this diversity makes machine learning algorithms difficult to generalize for such tasks, especially across multiple  ...  with weakly-supervised built-up area detection.  ... 
arXiv:2007.02277v1 fatcat:hoqvltpfabdzbcfujnnoyvtdum
« Previous Showing results 1 — 15 out of 2,943 results