605 Hits in 4.2 sec

Un-Mix: Rethinking Image Mixtures for Unsupervised Visual Representation Learning [article]

Zhiqiang Shen and Zechun Liu and Zhuang Liu and Marios Savvides and Trevor Darrell and Eric Xing
2022 arXiv   pre-print
The recently advanced unsupervised learning approaches use the siamese-like framework to compare two "views" from the same image for learning representations.  ...  Despite its conceptual simplicity, we show empirically that with the solution -- Unsupervised image mixtures (Un-Mix), we can learn subtler, more robust and generalized representations from the transformed  ...  Acknowledgement We thank all reviewers for their constructive and helpful comments in reviewing our paper. This manuscript has been revised significantly over its previous version.  ... 
arXiv:2003.05438v5 fatcat:njjxu2alp5hizbglqq35uidyee

Multi-task Mid-level Feature Alignment Network for Unsupervised Cross-Dataset Person Re-Identification [article]

Shan Lin, Haoliang Li, Chang-Tsun Li, Alex Chichung Kot
2018 arXiv   pre-print
To overcome this limitation, we develop a novel unsupervised Multi-task Mid-level Feature Alignment (MMFA) network for the unsupervised cross-dataset person re-identification task.  ...  In this way, the learned feature representation can be better generalised from one dataset to another which further improve the person re-identification accuracy.  ...  In our work, we rethink the assumption made for the unsupervised cross-dataset Re-ID.  ... 
arXiv:1807.01440v2 fatcat:26gw74g7nfblvf53edk32dc4ju

Unsupervised Foreground Extraction via Deep Region Competition [article]

Peiyu Yu, Sirui Xie, Xiaojian Ma, Yixin Zhu, Ying Nian Wu, Song-Chun Zhu
2021 arXiv   pre-print
In this work, we rethink the foreground extraction by reconciling energy-based prior with generative image modeling in the form of Mixture of Experts (MoE), where we further introduce the learned pixel  ...  We present Deep Region Competition (DRC), an algorithm designed to extract foreground objects from images in a fully unsupervised manner.  ...  These deficiencies ne- cessitate rethinking the problem of unsupervised foreground extraction.  ... 
arXiv:2110.15497v3 fatcat:kvb5x2j6s5cudksron43b424qy

Unsupervised Embedding Learning from Uncertainty Momentum Modeling [article]

Jiahuan Zhou, Yansong Tang, Bing Su, Ying Wu
2021 arXiv   pre-print
Existing popular unsupervised embedding learning methods focus on enhancing the instance-level local discrimination of the given unlabeled images by exploring various negative data.  ...  Moreover, the shortage of positive data and disregard for global discrimination consideration also pose critical issues for unsupervised learning but are always ignored by existing methods.  ...  [32] proposed a large-margin Gaussian Mixture (LGM) loss for image classification under the assumption that the deep features of the learning data follow a Gaussian Mixture distribution.  ... 
arXiv:2107.08892v1 fatcat:tkib2zoq7rci7ox7yjxiud6vpi

End-to-end Learning of a Fisher Vector Encoding for Part Features in Fine-grained Recognition [article]

Dimitri Korsch, Paul Bodesheim, Joachim Denzler
2020 arXiv   pre-print
We assume that part-based methods suffer from a missing representation of local features, which is invariant to the order of parts and can handle a varying number of visible parts appropriately.  ...  Part-based approaches for fine-grained recognition do not show the expected performance gain over global methods, although being able to explicitly focus on small details that are relevant for distinguishing  ...  In: Learning in Graphical Models, pp. 355-368. Springer (1998) 23. Perronnin, F., Dance, C.: Fisher kernels on visual vocabularies for image categorization.  ... 
arXiv:2007.02080v1 fatcat:f6n6rthtizcpphcldc6dlp7puq

SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decomposition [article]

Rishabh Kabra, Daniel Zoran, Goker Erdogan, Loic Matthey, Antonia Creswell, Matthew Botvinick, Alexander Lerchner, Christopher P. Burgess
2021 arXiv   pre-print
Leveraging the shared structure that exists across different scenes, our model learns to infer two sets of latent representations from RGB video input alone: a set of "object" latents, corresponding to  ...  We present an unsupervised variational approach to this problem.  ...  Space: Unsupervised object-oriented scene representation via spatial attention and decomposition. In International Conference on Learning Representations, 2020.  ... 
arXiv:2106.03849v2 fatcat:vtgeigfcrbgj5jnwka2e2djgxe

2021 Index IEEE Transactions on Image Processing Vol. 30

2021 IEEE Transactions on Image Processing  
The Author Index contains the primary entry for each item, listed under the first author's name.  ...  ., +, TIP 2021 1332-1341 DS-UI: Dual-Supervised Mixture of Gaussian Mixture Models for Uncer-tainty Inference in Image Recognition.  ...  Chen, J., +, TIP 2021 9164-9178 DS-UI: Dual-Supervised Mixture of Gaussian Mixture Models for Uncertainty Inference in Image Recognition.  ... 
doi:10.1109/tip.2022.3142569 fatcat:z26yhwuecbgrnb2czhwjlf73qu

Anomaly Clustering: Grouping Images into Coherent Clusters of Anomaly Types [article]

Kihyuk Sohn, Jinsung Yoon, Chun-Liang Li, Chen-Yu Lee, Tomas Pfister
2021 arXiv   pre-print
We compute weights in an unsupervised way or in a semi-supervised way if labeled normal data is available.  ...  Unlike object-centered image clustering applications, anomaly clustering is particularly challenging as anomalous patterns are subtle and local.  ...  Learning by trans- Matthijs Douze. Deep clustering for unsupervised learning duction. In Proceedings of the Fourteenth conference on Un- of visual features.  ... 
arXiv:2112.11573v1 fatcat:swvhwaq6tjcppo77wg2qtmi5bq

Anchoring to Exemplars for Training Mixture-of-Expert Cell Embeddings [article]

Siqi Wang, Manyuan Lu, Nikita Moshkov, Juan C. Caicedo, Bryan A. Plummer
2021 arXiv   pre-print
We propose Treatment ExemplArs with Mixture-of-experts (TEAMs), an embedding learning approach that learns a set of experts that are specialized in capturing technical variations in our training set and  ...  equipments used to collect microscopy images.  ...  Nature Re- Learning unsupervised feature representations for single cell views Drug Discovery, 20(2):145–159, 2021. 1 microscopy images with paired cell inpainting  ... 
arXiv:2112.03208v1 fatcat:dem557uqwjavboekwtwimk66hi

A Review of Open-World Learning and Steps Toward Open-World Learning Without Labels [article]

Mohsen Jafarzadeh, Akshay Raj Dhamija, Steve Cruz, Chunchun Li, Touqeer Ahmad, Terrance E. Boult
2022 arXiv   pre-print
and use them to define seven baselines for performance on the open-world learning without labels problem.  ...  , incremental learning, and instance management to learn new classes from a stream of unlabeled data in an unsupervised manner, survey how to adapt a few state-of-the-art techniques to fit the framework  ...  Efficientnet: Rethinking model scaling for convolutional neural networks.  ... 
arXiv:2011.12906v3 fatcat:psf56evv2vg63jaqwpzwgjwqhy

Unsupervised Domain Adaptation in Semantic Segmentation via Orthogonal and Clustered Embeddings [article]

Marco Toldo, Umberto Michieli, Pietro Zanuttigh
2020 arXiv   pre-print
Deep learning frameworks allowed for a remarkable advancement in semantic segmentation, but the data hungry nature of convolutional networks has rapidly raised the demand for adaptation techniques able  ...  Furthermore, we introduce two novel learning objectives to enhance the discriminative clustering performance: an orthogonality loss forces spaced out individual representations to be orthogonal, while  ...  A simple framework for contrastive learning of visual representations.  ... 
arXiv:2011.12616v1 fatcat:th6obcmtovcyjbxjqvgthtrtda

Unsupervised Myocardial Segmentation for Cardiac BOLD

Ilkay Oksuz, Anirban Mukhopadhyay, Rohan Dharmakumar, Sotirios A. Tsaftaris
2017 IEEE Transactions on Medical Imaging  
Our unsupervised approach outperforms even supervised state-of-the-art segmentation techniques by at least 10% when using Dice to measure accuracy on BOLD data and performs at-par for standard CINE MR.  ...  In this paper we build a joint motion and appearance method that relies on dictionary learning to find a suitable subspace.  ...  Texture information is generally considered as an effective feature representation of the myocardium for standard CINE MR images [21] .  ... 
doi:10.1109/tmi.2017.2726112 pmid:28708550 pmcid:PMC5726889 fatcat:ecxdykm2qzcvnahgxxtsjy5llq

Solving Inefficiency of Self-supervised Representation Learning [article]

Guangrun Wang, Keze Wang, Guangcong Wang, Philip H.S. Torr, Liang Lin
2021 arXiv   pre-print
Self-supervised learning (especially contrastive learning) has attracted great interest due to its huge potential in learning discriminative representations in an unsupervised manner.  ...  Despite the acknowledged successes, existing contrastive learning methods suffer from very low learning efficiency, e.g., taking about ten times more training epochs than supervised learning for comparable  ...  As is proved by [16, 43, 45, 57, 42] , unsupervised representation learning is critical to visual matching, therefore validating the effectiveness of our approach in re-ID is nontrivial.  ... 
arXiv:2104.08760v3 fatcat:prxoguwyzbh5jcacmo6n2wtauq

Rethinking the Value of Labels for Improving Class-Imbalanced Learning [article]

Yuzhe Yang, Zhi Xu
2020 arXiv   pre-print
We identify a persisting dilemma on the value of labels in the context of imbalanced learning: on the one hand, supervision from labels typically leads to better results than its unsupervised counterparts  ...  Real-world data often exhibits long-tailed distributions with heavy class imbalance, posing great challenges for deep recognition models.  ...  Unsupervised learning of visual representations by solving jigsaw puzzles. In European Conference on Computer Vision, pages 69-84. Springer, 2016.  ... 
arXiv:2006.07529v2 fatcat:m7vixcc6dzbvnh7oawwihff5u4

CryoDRGN: Reconstruction of heterogeneous structures from cryo-electron micrographs using neural networks [article]

Ellen D. Zhong, Tristan Bepler, Bonnie Berger, Joseph H Davis
2020 bioRxiv   pre-print
Here, we present cryoDRGN, an algorithm that for the first time leverages the representation power of deep neural networks to efficiently reconstruct highly heterogeneous complexes and continuous trajectories  ...  This ability enables cryoDRGN to discover previously overlooked structural states and to visualize molecules in motion.  ...  cryo-EM datasets 21 . 32 To learn this representation, cryoDRGN introduces an image-encoder/volume-decoder 33 framework to learn a latent representation of heterogeneity from single particle cryo-electron  ... 
doi:10.1101/2020.03.27.003871 fatcat:dhm65gstofg43feb64iymgkpgi
« Previous Showing results 1 — 15 out of 605 results