Deep Neural Networks Segment Neuronal Membranes in Electron Microscopy Images

By Dan Claudiu Ciresan, Alessandro Giusti, Luca Maria Gambardella, J. Urgen Schmidhuber. 2012

1,517 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

GloVe: Global Vectors for Word Representation (2014, background): “On the other hand, there is evidence that for certain types of neural networks, training multiple instances of the network and then combining the results can help reduce overﬁtting and noise and generally improve results (Ciresan et al., 2012).”
Fully convolutional networks for semantic segmentation (2014, methodology): “Patchwise training is common [27, 2, 7, 28, 9], but lacks the efficiency of fully convolutional training.”
U-Net: Convolutional Networks for Biomedical Image Segmentation (2015, methodology): “[2], whose best submission had a warping error of 0.”
Fast image scanning with deep max-pooling convolutional neural networks (2013, methodology): “detection [16].”
Evidential Deep Learning to Quantify Classification Uncertainty (2018, background): “Further ramifications that adapt neural nets to particular applications have brought unprecedented prediction accuracies, which in certain cases exceed human-level performance [5, 4].”
Backdooring Convolutional Neural Networks via Targeted Weight Perturbations (2018, background): “Convolutional Neural Networks (CNNs) have rapidly exceeded previous performance on a myriad of tasks [1, 4, 12, 14, 27], with image recognition being one of the most prominent.”
Deep learning in neural networks: An overview (2014, background): “2012 also saw the first pure image segmentation contest won by DL, again through an MC-GPUMPCNN (Segmentation of Neuronal Structures in EM Stacks Challenge, 2012; Ciresan et al., 2012a).”
Generating Adversarial Examples with Adversarial Networks (2018, background): “lications ranging from image recognition (Krizhevsky et al., 2012; He et al., 2016) to speech processing (Hinton et al., 2012) and from robotics training (Levine et al., 2016) to medical diagnostics (Ciresan et al., 2012). However, recent work has demonstrated that DNNs are vulnerable to adversarial”
Robust Physical-World Attacks on Deep Learning Models (2017, background): “Deep Neural Networks (DNNs) are being applied with great success to a variety of areas ranging from speech processing [6] to medical diagnostics [4] and are increasingly being used to control physical objects such as cars [14], UAVs [2], and robots [30].”
Explaining nonlinear classification decisions with deep Taylor decomposition (2015, background): “Recently, a particular class of nonlinear methods, Deep Neural Networks (DNNs), revolutionized the field of automated image classification by demonstrating impressive performance on large benchmark data sets [10, 11, 12].”
NAS-Unet: Neural Architecture Search for Medical Image Segmentation (2019, background): “It is also noted in [28] that segmenting an entire stackwith that approach is time-inefficiency, which takes at least 10 minutes for an stack on four GPUs.”
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].”
FlowNet: Learning Optical Flow with Convolutional Networks (2015, background): “class label) for each input image patch [8, 11].”
Evaluating the Visualization of What a Deep Neural Network Has Learned (2015, neutral): “Deep Neural Networks (DNNs) are powerful methods for solving large scale real world problems such as automated image classification [1], [2], [3], [4], natural language processing [5], [6], human action recognition [7], [8], or physics [9]; see”
Deep Networks with Internal Selective Attention through Feedback Connections (2014, background): “Deep convolutional neural networks (CNNs) [1] with max-pooling layers [2] trained by backprop [3] on GPUs [4] have become the state-of-the-art in object recognition [5, 6, 7, 8], segmentation/detection [9, 10], and scene parsing [11, 12] (for an extensive review see [13]).”
A survey on deep learning in medical image analysis (2017, methodology): “One of the earliest papers covering medical image segmentation with deep learning algorithms used such a strategy and was published by Ciresan et al. (2012).”
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.”
Indoor Semantic Segmentation using depth information (2013, background): “Training networks using multiscales representation appeared also the same year in [3, 21].”
N4-Fields: Neural Network Nearest Neighbor Fields for Image Transforms (2014, background): “CNNs are shown [2, 23, 3] to perform better if outputs of multiple models are averaged.”