A committee of neural networks for traffic sign classification

By Dan Claudiu Ciresan, Ueli Meier, Jonathan Masci, J. Urgen Schmidhuber. 2011

396 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Human-Computer Interaction

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Human-Computer Interaction

Citations in the corpus

Deep learning in neural networks: An overview (2014, neutral): “Currently successful techniques: LSTM RNNs and GPU-MPCNNs Most competition-winning or benchmark record-setting Deep Learners actually use one of two supervised techniques: (a) recurrent LSTM (1997) trained by CTC (2006) (Sections 5.13, 5.17, 5.21, 5.22), or (b) feedforward GPU-MPCNNs (2011, Sections”
Unmasking Clever Hans predictors and assessing what machines really learn (2019, background): “They occasionally exceed human performance spectacularly in tasks as diverse as face recognition [54], traffic sign classification [55], reading subway plans [56], understanding quantum many-body systems [43,44,47,57] or playing games such as Atari 2600 games [5,6], Go [7,8], Texas hold’em poker [10”
Deep Learning: Methods and Applications (2014, methodology): “Nevertheless, CNNs have been found highly effective and been commonly used in computer vision and image recognition [54, 55, 56, 57, 69, 198, 209, 212, 434].”
Convolutional neural networks applied to house numbers digit classification (2012, neutral): “In contrast, ConvNets learn features all the way from pixels to the classiﬁer.”
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models (2015, background): “, MNIST (2011), won numerous competitions [247, 38, 41, 39, 161, 42, 36, 134, 322, 37, 245], and achieved the first human-competitive or even superhuman results on well-known benchmarks, e.”
Self-Delimiting Neural Networks (2012, background): “(4) Online German Trafﬁc Sign Recognition Contest (1st & 2nd rank; 1.02% error rate) [8].”
Pedestrian detection with a Large-Field-Of-View deep network (2015, background): “Recently, deep learning methods have become the top performing approaches for pedestrian detection [10], [11], [12], and have also shown overwhelmingly good results in other computer vision applications [13], [14], [15], [16], and in other domains [17], [18], [19].”
Traffic-Sign Detection and Classification in the Wild (2016, methodology): “We have trained two CNNs for detecting traffic signs , and simultaneously detecting and classifying trafficsigns .”
Deep learning in robotics: a review of recent research (2017, background): “Many subsequent efforts built on these, but widespread interest in convolutional layers surged around 2012, when Krizhevsky used them to dominate in the ImageNet image recognition competition [41], and they were able to achieve super-human recognition on other notable image recognition benchmarks an”
Towards Explainable Artificial Intelligence (2019, background): “They not only outperform humans in complex visual tasks [16, 53] or strategic games [56, 83, 61], but also became an indispensable part of our every day lives, e.”
Multi-column deep neural network for traffic sign classification (2012, methodology): “Here we give a brief description of our architecture; a detailed description of the GPU implementation can be found in Ciresan et al. (2011a).”
Steel defect classification with Max-Pooling Convolutional Neural Networks (2012, neutral): “recognition benchmarks [5], [21], [22].”
Learning Hierarchical Features for Scene Labeling (2013, background): “low-level feature learning in the works of [42] and [6].”
A deep learning approach for Parkinson’s disease diagnosis from EEG signals (2018, methodology): “The convolutional layer convolves with the input signal using a kernel (window) [25, 26].”
Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognition (2012, methodology): “The approach in Ciresan et al. (2011) won the first stage of the GT- SRB competition by using a committee of CNNs trained on raw image pixels and multi-layer perceptrons (MLP) trained on the three provided HOG feature sets.”