Unsupervised Visual Representation Learning by Graph-Based Consistent Constraints

Li, Dong; Hung, Wei Chih; Huang, Jia; Wang, Shengjin; Ahuja, Narendra; Yang, Ming–Hsuan

doi:10.1007/978-3-319-46493-0_41

Cited by 56 publications

(41 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work of Li et al [30] is similar to ours in following a graph-based mining approach. In our comparisons, we show that choosing hard examples is essential and results in better performance for our approach.…”

Section: Related Workmentioning

confidence: 77%

Mining on Manifolds: Metric Learning Without Labels

İşcen

Tolias

Avrithis

et al. 2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

107

View full text Add to dashboard Cite

In this work we present a novel unsupervised framework for hard training example mining. The only input to the method is a collection of images relevant to the target application and a meaningful initial representation, provided e.g. by pre-trained CNN. Positive examples are distant points on a single manifold, while negative examples are nearby points on different manifolds. Both types of examples are revealed by disagreements between Euclidean and manifold similarities. The discovered examples can be used in training with any discriminative loss.The method is applied to unsupervised fine-tuning of pretrained networks for fine-grained classification and particular object retrieval. Our models are on par or are outperforming prior models that are fully or partially supervised.

show abstract

Section: Related Workmentioning

confidence: 77%

Mining on Manifolds: Metric Learning Without Labels

İşcen

Tolias

Avrithis

et al. 2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

107

View full text Add to dashboard Cite

show abstract

“…Compared with previous works, the top-1 retrieval accuracy is 0.659 while DenseSIFT and MVMME are below 0. 6. In order to demonstrate in details, we take DCIS and Invasive as example.…”

Section: Resultsmentioning

confidence: 99%

“…Starting from anchor node V As defined in [6], if V i belongs to its own n−order k−nearest neighbors, we can obtain a directed cycle as follows:…”

Section: Rs Imentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised Feature Learning for Endomicroscopy Image Retrieval

Vyas

Yang

et al. 2017

Medical Image Computing and Computer Assisted Intervention − MICCAI 2017

View full text Add to dashboard Cite

Abstract. Learning the visual representation for medical images is a critical task in computer-aided diagnosis. In this paper, we propose Unsupervised Multimodal Graph Mining (UMGM) to learn the discriminative features for probe-based confocal laser endomicroscopy (pCLE) mosaics of breast tissue. We build a multiscale multimodal graph based on both pCLE mosaics and histology images. The positive pairs are mined via cycle consistency and the negative pairs are extracted based on geodetic distance. Given the positive and negative pairs, the latent feature space is discovered by reconstructing the similarity between pCLE and histology images. Experiments on a database with 700 pCLE mosaics demonstrate that the proposed method outperforms previous works on pCLE feature learning. Specially, the top-1 accuracy in an eight-class retrieval task is 0.659 which leads to 10% improvement compared with the state-of-the-art method.

show abstract

“…The network transforms a classification network to multiple branches with shared parameters. Such network structures have been used for re-identification problems [31,1] and unsupervised visual learning [9,29,17]. Figure 2 shows the overview of the proposed algorithm.…”

Section: Related Workmentioning

confidence: 99%

Scene Parsing with Global Context Embedding

Hung

Tsai

Shen

et al. 2017

2017 IEEE International Conference on Computer Vision (ICCV)

Self Cite

View full text Add to dashboard Cite

We present a scene parsing method that utilizes global context information based on both the parametric and nonparametric models. Compared to previous methods that only exploit the local relationship between objects, we train a context network based on scene similarities to generate feature representations for global contexts. In addition, these learned features are utilized to generate global and spatial priors for explicit classes inference. We then design modules to embed the feature representations and the priors into the segmentation network as additional global context cues. We show that the proposed method can eliminate false positives that are not compatible with the global context representations. Experiments on both the MIT ADE20K and PASCAL Context datasets show that the proposed method performs favorably against existing methods.

show abstract

Unsupervised Visual Representation Learning by Graph-Based Consistent Constraints

Cited by 56 publications

References 37 publications

Mining on Manifolds: Metric Learning Without Labels

Mining on Manifolds: Metric Learning Without Labels

Unsupervised Feature Learning for Endomicroscopy Image Retrieval

Scene Parsing with Global Context Embedding

Contact Info

Product

Resources

About