Corpus ID: 231846995

Sill-Net: Feature Augmentation with Separated Illumination Representation

@article{Zhang2021SillNetFA,
  title={Sill-Net: Feature Augmentation with Separated Illumination Representation},
  author={Hanwang Zhang and Zhong Cao and Ziang Yan and Changshui Zhang},
  journal={ArXiv},
  year={2021},
  volume={abs/2102.03539}
}
For visual object recognition tasks, the illumination variations can cause distinct changes in object appearance and thus confuse the deep neural network based recognition models. Especially for some rare illumination conditions, collecting sufficient training samples could be time-consuming and expensive. To solve this problem, in this paper we propose a novel neural network architecture called Separating-Illumination Network (Sill-Net). Sill-Net learns to separate illumination features from… Expand
2 Citations

Figures and Tables from this paper

Squeezing Backbone Feature Distributions to the Max for Efficient Few-Shot Learning
TLDR
A novel transfer-based method which aims at processing the feature vectors so that they become closer to Gaussian-like distributions, resulting in increased accuracy in transductive few-shot learning. Expand
iELMNet: An Application for Traffic Sign Recognition using CNN and ELM
Traffic Sign Recognition (TSR) is a crucial step for automated vehicles and driver assistance systems. Automated TSD in an extreme environment has always been challenging due to foggy, rainy, blurry,Expand

References

SHOWING 1-10 OF 36 REFERENCES
Very Deep Convolutional Networks for Large-Scale Image Recognition
TLDR
This work investigates the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting using an architecture with very small convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. Expand
Dynamic Few-Shot Visual Learning Without Forgetting
TLDR
This work proposes to extend an object recognition system with an attention based few-shot classification weight generator, and to redesign the classifier of a ConvNet model as the cosine similarity function between feature representations and classification weight vectors. Expand
Deep Learning Logo Detection with Data Expansion by Synthesising Context
TLDR
This work designs a novel algorithm for generating Synthetic Context Logo (SCL) training images to increase model robustness against unknown background clutters, resulting in superior logo detection performance. Expand
Decoupling Representation and Classifier for Noisy Label Learning
TLDR
It is discovered that the representation is much more fragile in the presence of noisy labels than the classifier, and a new method, i.e., REED, is designed to leverage above discoveries to learn from noisy labels robustly. Expand
ImageNet Large Scale Visual Recognition Challenge
TLDR
The creation of this benchmark dataset and the advances in object recognition that have been possible as a result are described, and the state-of-the-art computer vision accuracy with human accuracy is compared. Expand
Dataset Augmentation in Feature Space
TLDR
This paper adopts a simpler, domain-agnostic approach to dataset augmentation, and works in the space of context vectors generated by sequence-to-sequence models, demonstrating a technique that is effective for both static and sequential data. Expand
STaDA: Style Transfer as Data Augmentation
TLDR
This work explores the state-of-the-art neural style transfer algorithms and applies them as a data augmentation method on Caltech 101 and Caltech 256 dataset, and finds around 2% improvement from 83% to 85% of the image classification accuracy with VGG16, compared with traditionalData augmentation strategies. Expand
Spatial Transformer Networks
TLDR
This work introduces a new learnable module, the Spatial Transformer, which explicitly allows the spatial manipulation of data within the network, and can be inserted into existing convolutional architectures, giving neural networks the ability to actively spatially transform feature maps. Expand
Deep Traffic Sign Detection and Recognition Without Target Domain Real Images
TLDR
A novel database generation method that requires only arbitrary natural images, i.e., requires no real image from the target-domain, and templates of the traffic signs is proposed, shown to be effective for the training of a deep detector on traffic signs from multiple countries. Expand
Co-domain Embedding using Deep Quadruplet Networks for Unseen Traffic Sign Recognition
TLDR
This work proposes a novel feature embedding scheme for unseen class classification when the representative class template is given, and performs co-domain embedding using a quadruple relationship from real and synthetic domains. Expand
...
1
2
3
4
...