Increasing robustness against background noise: Visual pattern recognition by a neocognitron

Fukushima, Kunihiko

doi:10.1016/j.neunet.2011.03.017

Cited by 21 publications

(8 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Different from the neocognitron of old versions, the inhibition from V-cells works, not in a divisional manner, but in a subtractive manner. This is effective for increasing robustness to background noise [5]. If inhibition works in a divisional manner, an S-cell, whose receptive field (left circle on U 0 ) covers only a part of the faint background character, responds strongly despite of the weak intensity of the stimulus.…”

Section: Robustness To Background Noisementioning

confidence: 99%

“…Recent neocognitrons use subtractive inhibition, while old neocognitrons used the divisional inhibition. (modified from[5])…”

mentioning

confidence: 99%

“…The recognition error under background noise, which is a faint image of a different digit. The recognition error remains small if the inhibition to S-cells works in a subtractive manner, but becomes very large if it works in a divisional manner[5].…”

mentioning

confidence: 99%

See 2 more Smart Citations

Recent advances in the deep CNN neocognitron

Fukushima

2019

NOLTA

Self Cite

View full text Add to dashboard Cite

Deep convolutional neural networks (deep CNN) show a large power for robust recognition of visual patterns. The neocognitron, which was first proposed by Fukushima (1979), is a network classified to this category. Its architecture was suggested by neurophysiological findings on the visual systems of mammals. It acquires the ability to recognize visual patterns robustly through learning. Although the neocognitron has a long history, improvements of the network are still continuing. This paper discusses the recent neocognitron, focusing on differences from the conventional deep CNN. Some other functions of the visual system can also be realized by networks extended from the neocognitron, for example, recognition of partly occluded patterns, the mechanism of selective attention, and so on.

show abstract

Section: Robustness To Background Noisementioning

confidence: 99%

“…Recent neocognitrons use subtractive inhibition, while old neocognitrons used the divisional inhibition. (modified from[5])…”

mentioning

confidence: 99%

See 1 more Smart Citation

Recent advances in the deep CNN neocognitron

Fukushima

2019

NOLTA

Self Cite

View full text Add to dashboard Cite

show abstract

“…Types of neural network, known as convolutional neural networks, had been trained to recognize patterns even if images are distorted to a degree [25], [26], [27], but taking the same pixel location of a scene over several snapshots and looking to see if it was cloudy or not is little information to use for making predictions. But vectors that show the motion of clouds have rich data about their direction and length (which show the direction and distance a cloud pixel travels).…”

Section: Neural Networkmentioning

confidence: 99%

Prediction of Cloud Movement from Satellite Images Using Neural Networks

Penteliuc

Frîncu

2019

2019 21st International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC)

View full text Add to dashboard Cite

Predicting cloud movement and dynamics is an important aspect in several areas, including prediction of solar energy generation. Knowing where a cloud will be or how it evolves over a given geographical area can help energy providers to better estimate their production levels. In this paper we propose a novel approach to predicting cloud movement based on satellite imagery. It combines techniques of generating motion vectors from sequential images with neural networks. First, the images are masked to isolate cloud pixels, then Farneback's version of the Optical Flow algorithm is used to detect motion from one image to the next and generate motion vector flow for each pair of images. After that, a feed forward back propagation neural network is trained with the vector data derived from the dataset imagery. Different parameters for the duration of the training, size of the input, and the neighborhood radius of one point in the scene are used. Promising results are presented and discussed to weight the potential of the proposed algorithm for forecasting cloud cover and cloud position in a scene.

show abstract

“…Обучение сверточных слоев в неокогнитроне производится с помощью локальных алгоритмов обучения без учителя, либо веса задаются заранее [45,46]. Для слоев подвыборки используется пространственное усреднение (spatial averaging) [43,47]. Таким образом, несмотря на то, что неокогнитрон является глубокой нейронной сетью, глубокое обучение в нем не используется.…”

Section: архитектура глубоких нейронных сетейunclassified

An Overview of Methods for Deep Learning in Neural Networks

2017

CMSE

View full text Add to dashboard Cite

Increasing robustness against background noise: Visual pattern recognition by a neocognitron

Cited by 21 publications

References 19 publications

Recent advances in the deep CNN neocognitron

Recent advances in the deep CNN neocognitron

Prediction of Cloud Movement from Satellite Images Using Neural Networks

An Overview of Methods for Deep Learning in Neural Networks

Contact Info

Product

Resources

About