Deep Rotation Equivariant Network

Li, Junying; Yang, Zichen; Liu, Haifeng; Cai, Deng

doi:10.1016/j.neucom.2018.02.029

Cited by 33 publications

(29 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dieleman et al () first tried the concept of Cyclic Pooling and Rolling on a different class of rotationally invariant images (galaxy morphology). An alternative method of obtaining rotational augmentations by Li et al () may be more efficient than the one we used from Lasagne (Dieleman et al ). Additionally, iterative augmentation is possible by using early models to iteratively process available unlabeled data in order to harvest additional training images, as described in Luo et al ().…”

Section: Discussionmentioning

confidence: 99%

Improving plankton image classification using context metadata

Ellen

Graff

Ohman

2019

Limnology & Ocean Methods

View full text Add to dashboard Cite

Advances in both hardware and software are enabling rapid proliferation of in situ plankton imaging methods, requiring more effective machine learning approaches to image classification. Deep Learning methods, such as convolutional neural networks (CNNs), show marked improvement over traditional feature-based supervised machine learning algorithms, but require careful optimization of hyperparameters and adequate training sets. Here, we document some best practices in applying CNNs to zooplankton and marine snow images and note where our results differ from contemporary Deep Learning findings in other domains. We boost the performance of CNN classifiers by incorporating metadata of different types and illustrate how to assimilate metadata beyond simple concatenation. We utilize both geotemporal (e.g., sample depth, location, time of day) and hydrographic (e.g., temperature, salinity, chlorophyll a) metadata and show that either type by itself, or both combined, can substantially reduce error rates. Incorporation of context metadata also boosts performance of the feature-based classifiers we evaluated: Random Forest, Extremely Randomized Trees, Gradient Boosted Classifier, Support Vector Machines, and Multilayer Perceptron. For our assessments, we use an original data set of 350,000 in situ images (roughly 50% marine snow and 50% nonsnow sorted into 26 categories) from a novel in situ Zooglider. We document asymptotically increasing performance with more computationally intensive techniques, such as substantially deeper networks and artificially augmented data sets. Our best model achieves 92.3% accuracy with our 27-class data set. We provide guidance for further refinements that are likely to provide additional gains in classifier accuracy.

show abstract

Section: Discussionmentioning

confidence: 99%

Improving plankton image classification using context metadata

Ellen

Graff

Ohman

2019

Limnology & Ocean Methods

View full text Add to dashboard Cite

show abstract

“…Recently there has been an explosion of interest into CNNs with predefined transformation equivariances, beyond translation [2,3,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19] [20, 21,22,23]. However, with the exception of Cohen and Welling [23] (projections on sphere), Kondor [22] (point clouds), and Thomas et al [20] (point clouds), these have mainly focused on the 2D scenario.…”

Section: Related Workmentioning

confidence: 99%

CubeNet: Equivariance to 3D Rotation and Translation

Worrall

Brostow

2018

Computer Vision – ECCV 2018

106

View full text Add to dashboard Cite

3D Convolutional Neural Networks are sensitive to transformations applied to their input. This is a problem because a voxelized version of a 3D object, and its rotated clone, will look unrelated to each other after passing through to the last layer of a network. Instead, an idealized model would preserve a meaningful representation of the voxelized object, while explaining the pose-difference between the two inputs. An equivariant representation vector has two components: the invariant identity part, and a discernable encoding of the transformation. Models that can't explain pose-differences risk "diluting" the representation, in pursuit of optimizing a classification or regression loss function.We introduce a Group Convolutional Neural Network with linear equivariance to translations and right angle rotations in three dimensions. We call this network CubeNet, reflecting its cube-like symmetry. By construction, this network helps preserve a 3D shape's global and local signature, as it is transformed through successive layers. We apply this network to a variety of 3D inference problems, achieving state-of-the-art on the ModelNet10 classification challenge, and comparable performance on the ISBI 2012 Connectome Segmentation Benchmark. To the best of our knowledge, this is the first 3D rotation equivariant CNN for voxel representations.

show abstract

“…(Fasel and Gatica-Perez 2006) and (Dieleman, Willett, and Dambre 2015) rotate the input itself before feeding it into stacks of CNNs and generating rotation invariant representations through gradual pooling or parameter sharing. (Teney and Hebert 2016;Wu, Hu, and Kong 2015;Li et al 2017) rotate the convolution filters (a cheaper albeit still expensive operation) instead of transforming the input followed by pooling. A similar approach was explored for scale by (Xu et al 2014).…”

Section: Prior Artmentioning

confidence: 99%

“…However, a number of studies have observed approximate albeit sufficient invariances in practice under this setting (Anselmi et al 2013;Pal, Juefei-Xu, and Savvides 2016;Pal et al 2017;Liao, Leibo, and Poggio 2013). The main motive for modelling transformations as unitary groups was to provide a theoretical connection to ConvNets and other methods that enforce other kinds of unitary invariance such as rotation invariance (Li et al 2017;Wu, Hu, and Kong 2015). However, real-world data experiences a large array of transformations acting, which certainly lie outside the span of unitary transformations.…”

Section: The Transformation Network Paradigmmentioning

confidence: 99%

“…Further, they can be optimized using vanilla gradient descent methods such as SGD. Unlike other methods that enforce additional invariances in convolutional architectures (Teney and Hebert 2016;Wu, Hu, and Kong 2015;Li et al 2017), NPTNs do not need to transform the input, activation maps or the filters at any stage of the learning/testing process. They enjoy benefits of a standard convolutional architecture such as speed and memory efficiency while being more powerful in modelling invariances and being elegant in their operation.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Non-Parametric Transformation Networks for Learning General Invariances from Data

Pal

Savvides

2019

AAAI

View full text Add to dashboard Cite

ConvNets, through their architecture, only enforce invariance to translation. In this paper, we introduce a new class of deep convolutional architectures called Non-Parametric Transformation Networks (NPTNs) which can learn general invariances and symmetries directly from data. NPTNs are a natural generalization of ConvNets and can be optimized directly using gradient descent. Unlike almost all previous works in deep architectures, they make no assumption regarding the structure of the invariances present in the data and in that aspect are flexible and powerful. We also model ConvNets and NPTNs under a unified framework called Transformation Networks (TN), which yields a better understanding of the connection between the two. We demonstrate the efficacy of NPTNs on data such as MNIST with extreme transformations and CIFAR10 where they outperform baselines, and further outperform several recent algorithms on ETH-80. They do so while having the same number of parameters. We also show that they are more effective than ConvNets in modelling symmetries and invariances from data, without the explicit knowledge of the added arbitrary nuisance transformations. Finally, we replace ConvNets with NPTNs within Capsule Networks and show that this enables Capsule Nets to perform even better.

show abstract

Deep Rotation Equivariant Network

Cited by 33 publications

References 27 publications

Improving plankton image classification using context metadata

Improving plankton image classification using context metadata

CubeNet: Equivariance to 3D Rotation and Translation

Non-Parametric Transformation Networks for Learning General Invariances from Data

Contact Info

Product

Resources

About