Artificial Immune Hybrid Photo Album Classifier
01 Jan 2017-pp 475-485
TL;DR: An Artificial Immune Hybrid Photo Album Classifier (AIHPAC) is proposed using the nonlinear biological properties of Human Immune Systems to develop an adaptive and automated personalized photo management system which efficiently manages and organizes personal photos.
Abstract: The personal photo collections are becoming significant in our day today existence. The challenge is to precisely intuit user’s complex and transient interests and to accordingly develop an adaptive and automated personalized photo management system which efficiently manages and organizes personal photos. This is increasingly gaining importance as it will be required to browse, search and retrieve efficiently the relevant information from personal collections which may extend from many years. Significance and relevance for the user also may undergo temporal and crucial shifts which need to be continually logged to generate patterns. The cloud paradigm makes available the basic platform but a system needs to be built wherein a personalized service with ability to capture diversity is guaranteed even when the training data size is small. An Artificial Immune Hybrid Photo Album Classifier (AIHPAC) is proposed using the nonlinear biological properties of Human Immune Systems. The system does event based clustering for an individual with embedded feature selection. The model is self learning and self evolving. The efficacy of the proposed method is efficiently demonstrated by the experimental results.
Citations
More filters
01 Jan 2020
TL;DR: This work proposes a novel framework where the small labelled dataset is appropriately augmented using the intelligent learning mechanisms of artificial immune systems to train the proposed model and shows that the generative deep framework utilizing artificial immune system principles provides a highly competitive approach for learning in the semi-supervised environment.
Abstract: Labelled data are not only time consuming but often expensive and difficult to procure as it involves skilful inputs by humans to tag and annotate. Contrary to this unlabelled data is comparatively easier to procure but fewer methods exist to optimally use them. Semi-Supervised Learning overcomes this problem and assists to build better classifiers by using unlabelled data along with sufficient labelled data and may actually yield higher accuracy with considerably less human input effort. But if the labelled data set is inadequate in size then the Semi-Supervised techniques are also stuck. We propose a novel framework where the small labelled dataset is appropriately augmented using the intelligent learning mechanisms of artificial immune systems to train the proposed model. The model retrains with the unlabelled data to fortify the learning mechanism. We show that the generative deep framework utilizing artificial immune system principles provides a highly competitive approach for learning in the semi-supervised environment.
References
More filters
09 Dec 2008
TL;DR: A Web-based intelligent photo browser which enables automatic clustering of unstructured personal digital photo collection and user studies show that APC is better in the case of the personal photographs while it requires much time for otherspsila photo collections.
Abstract: We developed a Web-based intelligent photo browser which enables automatic clustering of unstructured personal digital photo collection. We conducted user studies to assess the usability of the developed photo browser (APC, automatic photo classifier) compared with an unstructured one. The user task adopted here was finding some of target pictures indicated by an experimenter from personal or somebody elsepsilas photo collections. The results show that APC is better in the case of the personal photographs while it requires much time for otherspsila photo collections. It was suggested that the glance of a photo browser should be considered according to whether the photographs had been taken by the user.
4 citations
Journal Article•
TL;DR: In order to avoid the one-sidedness problems which may be produced by one single classifier, the multi-fusion method based on SVM-RKNN is presented.
Abstract: When Support Vector Machine(SVM) is used to solve the classification problems,the samples nearby the SVM hyperplanes are more easily misclassified.To solve this problem,the Reverse K-Nearest Neighbor method is introduced into the classification problems,and the Reverse K-Nearest Neighbor classification method(RKNN) is presented.And then,a new classification algorithm based on Support Vector Machine and Reverse K-Nearest Neighbor classification method(SVM-RKNN) is presented.At last,in order to avoid the one-sidedness problems which may be produced by one single classifier,the multi-fusion method based on SVM-RKNN is presented.The experimental results show that the average forecast accuracy of the SVM-RKNN method increases 2.13% than the SVM method,and the average forecast accuracy of the multi-fusion method based on SVM-RKNN increases 2.54% and 0.41% than the SVM and SVM-RKNN method respectively.
3 citations
Journal Article•
TL;DR: A dada-driven classification algorithm based on organizational evolution and entropy(DDCAOEE) was proposed, which uses a bottom-up search mechanism and can avoid generating meaningless rules during the evolutionary process.
Abstract: A dada-driven classification algorithm based on organizational evolution and entropy(DDCAOEE)was proposed.Different from the available evolutionary algorithms,the DDCAOEE uses a bottom-up search mechanism and this method can avoid generating meaningless rules during the evolutionary process.The organizations of information systems were constructed,and three evolutionary operators and a selection mechanism based on entropy were presented.An evolutionary method was devised for determining the significance of each attribute,and the fitness function for organizations was defined on the basis of significance.DDCAOEE was compared with two well-known classification algorithms,and the simulation results show that DDCAOEE achieves a higher predictive accuracy and a smaller number of rules.
3 citations
29 May 2012
TL;DR: A supervised classification algorithm based on artificial immune, which has non-linear and clone selection, immune regulation, immune memory and other features of biological immune system, which provides a new solution for supervised classification problem.
Abstract: In order to explore more efficient classification method, this paper presents a supervised classification algorithm based on artificial immune. It describes the representation of antibody and antigen in the classification algorithm, mathematical model of antibody population reproduction and immune memory formation. The experimental results show that the algorithm can achieve high classification performance. The average classification accuracy is 89.3%, stable classification performance. It has non-linear and clone selection, immune regulation, immune memory and other features of biological immune system, which provides a new solution for supervised classification problem.
1 citations