5 Hits in 2.4 sec

DeepUSPS: Deep Robust Unsupervised Saliency Prediction With Self-Supervision [article]

Duc Tam Nguyen, Maximilian Dax, Chaithanya Kumar Mummadi, Thi Phuong Nhung Ngo, Thi Hoai Phuong Nguyen, Zhongyu Lou, Thomas Brox
2021 arXiv   pre-print
These labels are refined incrementally in multiple iterations via our proposed self-supervision technique.  ...  In this work, we propose a two-stage mechanism for robust unsupervised object saliency prediction, where the first stage involves refinement of the noisy pseudo labels generated from different handcrafted  ...  DeepUSPS: Deep Unsupervised saliency prediction via self-supervision In this section, we explain the technical details of components in the overall pipeline shown in Fig. 3 .  ... 
arXiv:1909.13055v4 fatcat:wdx3l53gczcsnbdk4t6ed366ae

Characterizing and Improving the Robustness of Self-Supervised Learning through Background Augmentations [article]

Chaitanya K. Ryali, David J. Schwab, Ari S. Morcos
2021 arXiv   pre-print
We also make progress in completely unsupervised saliency detection, in the process of generating saliency masks used for background augmentations.  ...  Recent progress in self-supervised learning has demonstrated promising results in multiple visual tasks.  ...  Unsupervised Saliency Detection: DeepUSPS 2 In order to train a completely unsupervised saliency detector, we build upon DeepUSPS , a recent state-of-the-art weakly supervised saliency detection method  ... 
arXiv:2103.12719v2 fatcat:qaxu7ypdmvcqjez6s5hx3s6slq

Object Segmentation Without Labels with Large-Scale Generative Models [article]

Andrey Voynov, Stanislav Morozov, Artem Babenko
2021 arXiv   pre-print
The recent rise of unsupervised and self-supervised learning has dramatically reduced the dependency on labeled data, providing effective image representations for transfer to downstream vision tasks.  ...  Namely, we show that recent unsupervised GANs allow to differentiate between foreground/background pixels, providing high-quality saliency masks.  ...  To compare with deep saliency detection models, we also add DeepUSPS (Nguyen et al., 2019) to the list of baselines.  ... 
arXiv:2006.04988v2 fatcat:nt3ae3lbsra3tbqwyu5stoce64

Distilling Localization for Self-Supervised Representation Learning [article]

Nanxuan Zhao, Zhirong Wu, Rynson W.H. Lau, Stephen Lin
2021 arXiv   pre-print
With this approach (DiLo), significant performance is achieved for self-supervised learning on ImageNet classification, and also for object detection on PASCAL VOC and MSCOCO.  ...  Recent progress in contrastive learning has revolutionized unsupervised representation learning.  ...  saliency prediction via selfsupervision.  ... 
arXiv:2004.06638v2 fatcat:36x43zh5vvg3pbsvak7wf3nwci

Few-Cost Salient Object Detection with Adversarial-Paced Learning [article]

Dingwen Zhang, Haibin Tian, Jungong Han
2021 arXiv   pre-print
A fundamental challenge in training the existing deep saliency detection models is the requirement of large amounts of annotated data.  ...  Essentially, APL is derived from the self-paced learning (SPL) regime but it infers the robust learning pace through the data-driven adversarial learning mechanism rather than the heuristic design of the  ...  In light of the advanced development in deep learning, recent salient object detection methods mainly adopted the CNN models to learn saliency patterns under a fully supervised fashion.  ... 
arXiv:2104.01928v1 fatcat:vrcarkcjtrbgxfiii4znjdqz34