Robust Object Co-detection

Guo, Xin; Liu, Dong; Jou, Brendan; Zhu, Mojun; Cai, Anni; Chang, Shih-Fu

doi:10.1109/cvpr.2013.412

Cited by 27 publications

(23 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To effectively utilize prior knowledge from a collection of labeled examples, we devise a structure pursuit model for semantic segmentation and labeling of 3D meshes based on low-rank representation (LRR), which is a powerful tool to recover the relationship between data [3][4][5]. Different from the meaning in computer graphics, the "structure" above means the underlying structure of the features.…”

Section: Introductionmentioning

confidence: 99%

Low-rank 3D mesh segmentation and labeling with structure guiding

Liu

Zhang

Liu

et al. 2015

Computers & Graphics

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Low-rank 3D mesh segmentation and labeling with structure guiding

Liu

Zhang

Liu

et al. 2015

Computers & Graphics

View full text Add to dashboard Cite

“…By contrast, object co-detection [2] attempts to simultaneously exploit the similarity between pairs of objects to perform detection jointly in multiple images. Most existing co-detection methods rely on simple handcrafted features to model object similarity [2,8,20,11]. By contrast, in our previous work [10], we performed feature selection using pre-trained CNN features.…”

Section: Related Workmentioning

confidence: 99%

“…By contrast, in a parallel line of research, object co-detection [2,8,10,11] has emerged as an effective approach to leveraging the information jointly contained in multiple images to improve detection accuracy. Unfortunately, to model the similarity of multiple objects, existing methods rely on either handcrafted features [2,8,20,11], or features learned for object recognition [10]. As a consequence, they are ill-suited to handle general object proposals, whose appearance is subject to much larger variations than specific object classes.…”

Section: Introductionmentioning

confidence: 99%

Learning to Co-Generate Object Proposals with a Deep Structured Network

Hayder

Salzmann

2016

2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

Generating object proposals has become a key component of modern object detection pipelines. However, most existing methods generate the object candidates independently of each other. In this paper, we present an approach to co-generating object proposals in multiple images, thus leveraging the collective power of multiple object candidates. In particular, we introduce a deep structured network that jointly predicts the objectness scores and the bounding box locations of multiple object candidates. Our deep structured network consists of a fully-connected Conditional Random Field built on top of a set of deep Convolutional Neural Networks, which learn features to model both the individual object candidates and the similarity between multiple candidates. To train our deep structured network, we develop an end-to-end learning algorithm that, by unrolling the CRF inference procedure, lets us backpropagate the loss gradient throughout the entire structured network. We demonstrate the effectiveness of our approach on two benchmark datasets, showing significant improvement over state-of-the-art object proposal algorithms.

show abstract

“…However these tend to require that the object be in each image considered and are as such not appropriate for our noisy annotation task. A recent method [15] could provide an alternative method of achieving visual consistency amongst the candidates.…”

Section: Searching For Candidate Regionsmentioning

confidence: 99%

Of Gods and Goats: Weakly Supervised Learning of Figurative Art

Crowley¹,

Zisserman²

2013

Procedings of the British Machine Vision Conference 2013

View full text Add to dashboard Cite

The objective of this paper is to automatically annotate images of gods and animals in decorations on classical Greek vases. Such images often require expert knowledge in labelling. We start from a large dataset of images of vases with associated brief text descriptions.We develop a weakly supervised learning approach to solve the correspondence problem between the descriptions and unknown image regions. The approach progressively strengthens the supervision so that eventually a Deformable Part Model (DPM) sliding window detector can be learnt (for each god/animal) and used to annotate all vases by detection. There are two key steps: first, text mining the vase descriptions to obtain clusters for each god where there is a visual consistency between at least a subset of the images; and second, discriminatively searching within these clusters for consistent regions that are used as positive training examples for the DPM.The method successfully annotates a large variety of Gods and other animals, and we include a quantitative evaluation over hundreds of images.

show abstract

Robust Object Co-detection

Cited by 27 publications

References 25 publications

Low-rank 3D mesh segmentation and labeling with structure guiding

Low-rank 3D mesh segmentation and labeling with structure guiding

Learning to Co-Generate Object Proposals with a Deep Structured Network

Of Gods and Goats: Weakly Supervised Learning of Figurative Art

Contact Info

Product

Resources

About