Convolutional Neural Networks and Texture Classification
Convolutional neural networks (CNN) model is an instrumental computational model not only in computer vision but also in many image and video applications. Similar toCognitron and Neocognitron, CNN can automatically learn the features of data with the mul
- PDF / 1,502,361 Bytes
- 19 Pages / 439.37 x 666.142 pts Page_size
- 92 Downloads / 265 Views
10
I have just three things to teach: simplicity, patience, compassion. These three are your greatest treasures. — Lao Tzu
Convolutional neural networks (CNN) model is an instrumental computational model not only in computer vision but also in many image and video applications. Similar to Cognitron and Neocognitron, CNN can automatically learn the features of data with the multiple layers of neurons in the network. There are several different versions of the CNN which have been reported in the literature. If an original image texture is fed into the CNN, it will be called an image-based CNN. A major problem with the image-based CNNs is that the number of training images is very demanding for the good generalization of the network due to the rotation and scaling change in images. An alternative method is to divide an image into many small patches for the CNN training. This is very similar to the patches used in the K-views model. In this chapter, we will briefly explain the image-based CNN and patch-based CNN for image texture classification. The LeNet-5 neural network architecture will be used as a basic CNN model. CNN is useful not only in the image recognition but also in the textural feature representation. Texture features, which are automatically learned and extracted from a massive amount of images using the CNN, become the focus of developing feature extraction methods.
10.1
Convolutional Neural Networks (CNN)
Deep machine learning is a trend in recent years for the general pattern recognition and machine vision. Most deep machine learning is based on the neural networks approach which uses many layers of neurons; it is more than the number of layers used in the traditional artificial neural networks. Hence, it is called “deep” neural networks. Due to the advance of high-speed computing devices, several high-performance deep neural network models based on the foundation of the © Springer Nature Switzerland AG 2019 C.-C. Hung et al., Image Texture Analysis, https://doi.org/10.1007/978-3-030-13773-1_10
233
234
10
Convolutional Neural Networks and Texture Classification
traditional neural networks have been proposed and widely used in the broad domain of artificial intelligence. One of the deep neural networks, which is called the convolutional neural networks (CNN), might be the most popular in applications [3, 6, 7, 16, 24]. Unlike most traditional pattern recognition systems which require engineer-designed feature extractors, the CNN can automatically learn the feature maps through the deep layers of the networks. This characteristic feature makes CNN suitable for the big data analysis in machine learning. Even more, some advanced CNN has been developed for understanding convolutional networks and generating image descriptions [12, 31]. Convolutional neural networks (CNN) has become a de facto neural network used in visual recognition recently. The architecture of CNN is more or less similar to that of Neocognitron which is one of the earliest deep neural network models although several features used
Data Loading...
