Download Perspectives on Content-Based Multimedia Systems by Jian Kang Wu, Mohan S. Kankanhalli, Joo-Hwee Lim, Dezhong PDF

By Jian Kang Wu, Mohan S. Kankanhalli, Joo-Hwee Lim, Dezhong Hong

Multimedia information comprising of pictures, audio and video is turning into more and more universal. The reducing expenditures of purchaser digital units corresponding to electronic cameras and electronic camcorders, besides the benefit of transportation facilitated through the web, has bring about an exquisite upward push within the quantity of multimedia information generated and allotted. provided that this pattern of elevated use of multimedia facts is probably going to speed up, there's an pressing desire for supplying a transparent technique of shooting, storing, indexing, retrieving, interpreting and summarizing such facts.
Content-based entry to multimedia facts is of fundamental value because it is the common method through which people engage with such details. To facilitate the content-based entry of multimedia info, step one is to derive characteristic measures from those facts in order that a characteristic house illustration of the knowledge content material might be shaped. this may accordingly enable for mapping the function house to the logo house (semantics) both instantly or via human intervention. hence, sign to image mapping, important for any sensible procedure, should be effectively completed.
Perspectives on Content-Based Multimedia Systems presents a complete set of ideas to take on those vital matters. This publication deals precise strategies to a variety of useful difficulties in development actual platforms by way of delivering specifics of 3 platforms outfitted by means of the authors. whereas delivering a platforms concentration, it additionally equips the reader with a willing realizing of the basic concerns, together with a formalism for content-based multimedia database platforms, multimedia function extraction, object-based suggestions, signature-based innovations and fuzzy retrieval suggestions. The functionality evaluate problems with functional platforms is additionally defined. This e-book brings jointly crucial components of creating a content-based multimedia database procedure in a fashion that makes them available to practitioners in laptop technological know-how and electric engineering. it may well additionally function a textbook for graduate-level courses.

Show description

Read Online or Download Perspectives on Content-Based Multimedia Systems PDF

Best structured design books

Transactions on Computational Systems Biology IX

The LNCS magazine Transactions on Computational structures Biology is dedicated to inter- and multidisciplinary study within the fields of laptop technological know-how and lifestyles sciences and helps a paradigmatic shift within the strategies from computing device and data technological know-how to deal with the recent demanding situations coming up from the platforms orientated perspective of organic phenomena.

Interactive Relational Database Design: A Logic Programming Implementation

Relational databases have quick end up considered as a typical and effective method of organizing details. replica information might be eradicated and strong set-theoretic operations can be utilized to govern information. yet discovering the suitable relatives for a database isn't really but a trivial step for the uninitiated.

Human Identification Based on Gait

Biometrics now have an effect on many people's lives, and is the point of interest of a lot educational learn and advertisement improvement. Gait is likely one of the most modern biometrics, with its personal specific merits. Gait acknowledges humans incidentally they stroll and run, analyzes movement,which in flip implies reading sequences of pictures.

Extra resources for Perspectives on Content-Based Multimedia Systems

Sample text

The difference between di and ti is treated as an error Ei = di – ti which can be used to modify the parameters in the classifier towards less erroneous classification performance. In our approach described in this chapter, the training patterns and objectives are quite different. First, the input vectors for learning are not feature vectors, but similarities (or distances) between the feature vectors of two multimedia objects. Second, the output vectors for learning are not desired and target class labels, but mis-rank information.

W. : Storage and retrieval for image and video databases, San Jose, USA, pp. 173- 181, 1993 8. A. W. Picard, S. : Storage and Retrieval of Image and Video Databases 11, Vol. 2185, pp. 6-10, 1994 9. S. Santini, R. Jain, “Similarity matching”, IEEE Trans. on Pattern Analysis and Machine Intelligence, 1998 10. K. al. “Content-Based Retrieval for Trademark Registration”, Multimedia Tools and Applications, Vol. 3, No. 3, pp. 245-267, 1996 11. K. Wu, “Recognition by Recall”, RWC Symp. 97. 48 Chapter 2 12.

K) is the desired retrieval sequence for the sample si in a proper order. ; ... Having compared the retrieval results with the ideal one, we found that some retrieval objects should appear more to the front, and some should move more to the rear. We refer these absolute position shifts as mis-rank. , ... If we assign a unit loss to a mis-rank, the total loss of the retrieval would be: m Σij ij /K L= K where K is the length of retrieval sequence, and the average loss of the retrieval would be: ^ _L L= N where N is the number of mis-rank sequences in the training data set.

Download PDF sample

Rated 4.37 of 5 – based on 19 votes