This is known as 'Curse of Dimensionality', which states that the number of examples necessary to reliable generalization grows exponentially with the number of dimensions. Learn ability necessitates dimensionality reduction, which is the process of reducing the number of random features under consideration during image retrieval (Roweis and Saul, 2000).
In large multimedia databases, high-dimensional representation is computationally intensive and most users are unwilling to wait for results for a long time. Thus, for storage and retrieval efficiency concerns, dimensionality reduction in CBIR systems is necessary. Example of these techniques includes Principal Component Analysis (PCA), Independent Component Analysis (ICA) and Linear Discriminat Analysis (LDA).
1.3.3. Similarity Measure
The similarity between two images (represented by their feature values) is defined by a similarity measure. It is a measure that reflects the degree of closeness or separation of the images under consideration. In general, similarity/distance measures map the distance or similarity between the two images into a single numeric value and depend on two factors (Torres et al., 2011) as given below.
(i) the properties of the two objects
(ii) the measure itself.
Selection of similarity metrics has a direct impact on the performance of CBIR systems. The kind of feature vectors selected determines the kind of measurement that will be used to compare their similarity. If the features extracted from the images are presented as multi-dimensional points, the distances between corresponding multi-dimensional points can be calculated. Euclidean distance is the most common metric used to measure the distance between two points in multi-dimensional space (...
... middle of paper ...
...ruction and Image Retrieval).
Chapter 7, Results and Discussion, tabulates and discusses the various results obtained while testing the proposed CBIR systems. The results obtained are analyzed and discussed in this chapter. The findings of the study are summarized along with future research directions in Chapter 8, Summary and Conclusion.
The work of several researchers are quoted and used as evidence to support the concepts explained in this dissertation. All such evidences used are listed in the Bibliography of the dissertation.
This chapter provided a brief introduction to image retrieval and general framework of CBIR systems. The objectives formulated were also outlined. To achieve the objectives outlined in this chapter, a review of the previous research work was studied and the scrutinized works are summarized in the next chapter, Review of Literature.
The research questions being proposed in this paper include but are not limited the following
Color is widely remarked as one of the most demonstrative visual features, and as such it has been largely studied in the context of CBIR, thus number one to a rich variety of descriptors. As traditional color features used in CBIR, there are color histogram, color correlogram, and dominant color descriptor (DCD) [1,3,4]. A simple color similarity between two images can be
In response to the question set, I will go into detail of the study, consisting of the background, main hypotheses, as well the aims, procedure and results gathered from the study; explaining the four research methods chosen to investigate, furthering into the three methods actually tested.
To test these hypotheses, we first collected a two point discriminator containing a variety of distances for two points. The point distances on the discriminator included values of.25 millimeters being the smallest, ...
Law of Similarity: the law of similarity suggests that similar things tend to appear grouped together.
From many points of views, it can be considered as the starting point. The team working with it has a dream to make more objects recognition which is context base. They also have a desire to make the recognition more interactive. A new and exceptional feature has been suggested where a particular part of an image can be tapped and the information can be heard.
This article hasn’t provided an introduction; however a lengthy summary of the study which identifies the problem, purpose and rationale for the research study has been provided in the background. The introduction should give the reader a general sense of what the document is about, and preferably persuade the reader to continue reading. This prepares the reader for reading the rest of the document (Burns & Grove, 2001 p.636; Nieswiadomy, 2008 p.380; Stockhausen and Conrick, 2002).
[4] Borut Batagelj, (May 2006). Face recognition in different subspaces - A comparative study”, 6th International Workshop on Pattern Recognition in Information Systems, PRIS 2006 in conjunction with ICEIS 2006, Paphos, Cyprus.
In the modern era known as the “Information Age,” forms of electronic information are steadily becoming more important. Unfortunately, maintenance of data requires valuable resources in storage and transmission, as even the presence of information in storage re-quires some power. However, some of the largest files are those that are in formats re-plete with repetition, and thus are larger than they need to be. The study of data compres-sion is the science which attempts to advance toward methods that can be applied to data in order to make it take up less space. The uses for this are vast, and algorithms will need to be improved in order to sustain the inevitably larger files of the future. Thus, I decided to focus my research on the techniques that successful methods use in order to save space. My research question:
[Jain, 2004] Jain, A.K.;Ross, A.;Prabhakar, S.;"An introduction to biometric recognition", Volume: 14 Issue: 1 Issue Date: Jan. 2004, on page(s): 4 - 20
It is generally easy for a person to differentiate the sound of a human voice, from that of a violin; a handwritten numeral "3," from an "8"; and the aroma of a rose, from that of an onion. However, it is difficult for a programmable computer to solve these kinds of perceptual problems. These problems are difficult because each pattern usually contains a large amount of information, and the recognition problems typically have an inconspicuous, high-dimensional, structure.
Then classification is performed on the basis of similarity score of a class with respect to a neighbor.
2. Face recognition: Face recognition is based on both the shape and location of the eyes, eyebrows, nose, lips and chin. It is non intrusive method and very popular also. Facial recognition is carried out in two ways ...
The distance where by a search scanner can discover something different first from its surroundings but not yet identify it.
[19] K. Burgers, et al., A Comparative Analysis of Dimension Reduction Algorithms on Hyperspectral Data, 2009