A Unique Strategy for Swift Generation and Contrast of Applied Feature Vectors
N. Navya Teja , Mahender Veshala, , ,
Affiliations
:
Abstract
The easiest formula for determining an example in the test set is known as the closest Neighbor method. The item of great interest is in comparison to each sample within the training set, utilizing a distance measure, a similarity measure, or a mix of measures. The conventional deviation, also is referred to as square cause of the variance, informs us something concerning the contrast. It describes multiplication within the data, so a higher contrast image has a high variance; along with a low contrast image have a low variance. Even though this techniques could be enhanced if some pre-processing steps are utilized. In content-based image retrieval systems (CBIR) the best and straightforward searches would be the color based searches. In CBIR image classification needs to be computationally fast and efficient. Within this paper a brand new approach is introduced, which works according to low-level image histogram features. The primary benefit of this process may be the extremely swift generation and comparison from the applied feature vectors. It also includes the analysis of pre-processing calculations and the look classification. We are able to result in the Nearest Neighbor method better quality by choosing not only the nearest sample within the training set, but also by thought on several close feature vectors. Using each training set, the histograms from the three color channels were produced and also the above mentioned histogram features were calculated.
Citation
N. Navya Teja et.al," A Unique Strategy for Swift Generation and Contrast of Applied Feature Vectorsâ€, International Journal of Computer Engineering In Research Trends, Volume 3, Issue 8, August-2016, pp. 431-435
Authors are not required to pay any article-processing charges (APC) for their article to be published open access in Journal IJCERT. No charge is involved in any stage of the publication process, from administrating peer review to copy editing and hosting the final article on dedicated servers. This is free for all authors.
News & Events
Latest issue :Volume 10 Issue 1 Articles In press
Key Dates
☞ INVITING SUBMISSIONS FOR THE NEXT ISSUE :
☞ LAST DATE OF SUBMISSION : 31st March 2023
☞ SUBMISSION TO FIRST DECISION : In 7 Days
☞ FINAL DECISION : IN 3 WEEKS FROM THE DAY OF SUBMISSION
Important Announcements
All the authors, conference coordinators, conveners, and guest editors kindly check their articles' originality before submitting them to IJCERT. If any material is found to be duplicate submission or sent to other journals when the content is in the process with IJCERT, fabricated data, cut and paste (plagiarized), at any stage of processing of material, IJCERT is bound to take the following actions.
1. Rejection of the article.
2. The author will be blocked for future communication with IJCERT if duplicate articles are submitted.
3. A letter regarding this will be posted to the Principal/Director of the Institution where the study was conducted.
4. A List of blacklisted authors will be shared among the Chief Editors of other prestigious Journals
We have been screening articles for plagiarism with a world-renowned tool: Turnitin However, it is only rejected if found plagiarized. This more stern action is being taken because of the illegal behavior of a handful of authors who have been involved in ethical misconduct. The Screening and making a decision on such articles costs colossal time and resources for the journal. It directly delays the process of genuine materials.