English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Considering that PCNN is in accordance with human vision, so it has a good performance in image binaryzation and segmentation, the PCNN-based methods have advantage in feature extraction and image classification. In this paper, we propose a new method for feature extraction and image classification based PCNN which is more robust than others. This method mainly analyzes the quality and centroid of the image and extracts features from the relationship between different centroid. Then, the Euclidean distance distribution parameters used to describe these features. In image classification, the feature parameter also play an important role and distinguishes the difference between various object images efficiently. Introduction Feature extraction, as the basic part of image recognition, classification, and many other topics, is an important issue in image processing and computer version. In image processing, feature extraction used to detect and isolate various desired portions or shape of a digital image. Many algorithm proposed for getting more robust and efficient features. In low-level feature extraction, the research mostly concentrates on several direction such as edge detection[1,2,3], corner detection[4,5], blob detection[6,7,8,9], ridge detection[10,11], and scale-invariant feature transform (SIFT)[12,13,14,15,16]. In addition, there are also many shaped-based method like, template matching[18,19,20], thresholding[17], and hough transform[21,22,23,24]. PCNN is a self-organizing network that does not require training and the network was constructed by simulating the activities of the mammal’s visual cortex neurons[38].Considering that PCNN is in accordance with human vision and has a good performance in image binaryzation and segmentation, the PCNN-based methods have advantage in feature extraction and image classification. Johnson proposed that using firing time as the features[25,26,27]. While, this kind of method would be badly effect by illumination change. In some researches, using the characteristics of PCNN entropy series, people extract the features and recognize some object in image [28,29]. What’s more, the conception of quality is introduced into the image processing field and normalized moment of inertia (NMI) is used for feature extraction[30]. In this paper, we proposed a new method for feature extraction based on PCNN. Similar to the NMI feature extraction, we also use the image quality in processing and the quality parameters would play an important role in describing distribution of different pixels in an image. But the NMI feature would tend to be the same when there is a long strips or a round object in the image without background. That means NMI feature is unable to distinguish between some particular shapes for example a coin and a pen. In the method we proposed, the feature depends on the distribution of many partial or global centroid, rather than only a number of rotational inertia of the whole image in NMI method. Besides, a series of Euclidean distance would be adapt to describe the 563 Advances in Computer Science Research (ACRS), volume 54 International Conference on Computer Networks and Communication Technology (CNCT2016)

A New Method for Feature Extraction and Image Classification Based on PCNN and Image Quality