Learning to Label Aerial Images from Noisy Data

By Volodymyr Mnih, Geoffrey E. Hinton. 2012

410 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Remote Sensing

Citations in the corpus

Advances and Open Problems in Federated Learning (2019, background): “parate entities control the server, clients, and network. Similar distortions can occur due to the client data. Even if the data on a client is not intentionally malicious, it may have noisy features [301] (eg. in vision applications, a client may have a low-resolution camera whose output is scaled ”
The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (2015, methodology): “4 is inspired by methods that propose to learn from web data [13, 8, 9, 60, 40, 29, 16] or reason about label noise [34, 57, 52, 47, 38].”
Training Deep Neural Networks on Noisy Labels with Bootstrapping (2014, background): “Mnih & Hinton (2012) developed deep neural networks for improved labeling of aerial images, with robust loss functions to handle label omission and registration errors.”
Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noise (2018, methodology): “Unlike Mnih and Hinton [14] who focus on binary classification of aerial images and Larsen et al. [8] who assume symmetric label noise, [17, 24] consider label noise in the multi-class problem setting with asymmetric noise.”
Learning from Noisy Labels with Deep Neural Networks (2014, methodology): “In [11, 9], noise modeling is incorporated to neural network in the same way as our proposed model.”
Robust Loss Functions under Label Noise for Deep Neural Networks (2017, methodology): “[19] use a generative model for label corruption, estimated through an approximate EM algorithm, and show its eﬀectiveness in a binary classiﬁcation problem.”
Learning from massive noisy labeled data for image classification (2015, background): “Noise modeling with deep learning: Various methods have been proposed to handle label noise in different problem settings, but there are very few works about deep learning from noisy labels [13, 18, 24].”
Robust Loss Functions for Learning Multi-class Classifiers (2018, background): “Mnith et al. [12] proposed a generative model for label corruption in binary images (binary classiﬁ-cation) with approximate EM for estimation of parameters.”
Submitted to Ieee Transactions on Pattern Analysis and Machine Intelligence 1 Multiscale Centerline Detection (neutral): “Some works therefore aim at segmenting linear structures in biomedical images [5], [17], [42], [59] or aerial ones [37], [57] by applying classification to label the pixels or voxels as belonging to the structure of interest or to the background.”
Making Deep Neural Networks Robust to Label Noise: A Loss Correction Approach (2016, background): “[27] builds a noise model for binary classiﬁcation of aerial image patches, which can handle omission and wrong location of training labels.”
Seeing through the Human Reporting Bias: Visual Classifiers from Noisy Human-Centric Labels (2015, background): “Explicitly modeling label noise has received increasing attention in recent years [35, 36, 47, 54].”
Osprey: Weak Supervision of Imbalanced Extraction Problems without Code (2019, background): “Prior work in weak supervision has focused on the setting of independent crowd workers [7, 9], custom-tailored and hand-tuned distant supervision strategies in the natural language processing domain [14, 21], knowledge-bases[12, 14], handling generic label noise or mis-specification [3, 15].”
YouTube-8M: A Large-Scale Video Classification Benchmark (2016, neutral): “8% precision) or missing [40, 25] ( 14 .”
Deep learning with noisy labels: exploring techniques and remedies in medical image analysis (2019, background): “[141] proposed a similar loss function based on penalizing the disagreement between the predicted label and the posterior of the true label.”
Topological Map Extraction From Overhead Images (2018, background): “To the best of our knowledge, the ﬁrst (non-convolutional) deep learning approach to road network extraction was proposed by [35, 36].”
DeepRoadMapper: Extracting Road Topology from Aerial Images (2017, methodology): “In [16], existing maps are used for data augmentation.”
From Facial Parts Responses to Face Detection: A Deep Learning Approach (2015, background): “This setting is widely employed in image labeling [5, 24].”
Snorkel DryBell: A Case Study in Deploying Weak Supervision at Industrial Scale (2018, background): “In response, many machine learning developers have increasingly turned to weaker methods of supervision, in which a larger volume of cheaper, noisier labels is used [6, 9, 10, 23, 24, 29, 40].”
A Light CNN for Deep Face Representation With Noisy Labels (2015, background): “Mnih and Hinton [24] introduce two robust loss functions for noisy label aerial images.”
Learning From Noisy Labels With Deep Neural Networks: A Survey (2020, background): “Hence, several convex surrogate loss functions, which approximate the 0–1 loss function, have been proposed to train a speciﬁed classiﬁer under the binary classiﬁcation setting [47]–[51].”