Streamline of Cross Media Retrieval Using Term Frequency-Inverse Document Frequency And Color Histogram
Monelli Ayyavaraiah
Monelli Ayyavaraiah, Assistant Professor, Department of CSE, Sai Rajeswari Institute of Technology- Proddatur (Andhra Pradesh), India.
Manuscript received on 18 April 2019 | Revised Manuscript received on 25 April 2019 | Manuscript published on 30 April 2019 | PP: 450-455 | Volume-8 Issue-4, April 2019 | Retrieval Number: D6390048419/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Cross media retrieval provides the different media results such as text, video and audio for the single query. Many researches has been carried out for the cross-media retrieval due to its benefits. In this research, the features selection method such as Term Frequency and Inverse document frequency with color histogram (TFIDF-CH) is proposed for the cross-media retrieval system. Wikipedia dataset is popular dataset for the cross-media retrieval method and this is used to test the function of the proposed method. The text and image from the database are represent in the Bag-of-Words (BoW) and Visual BoW respectively. The TF-IDF feature is extracted from the text and color histogram is selected from the images. The graph is drawn based on these features stored in the dictionary learning. Then applied Minkowski distance to calculate the similarities between the different media. The TFIDF-CH achieves the average Mean Average Precision (MAP of 59.025 % compared with existing method having average MAP of 41.32%.
Keywords: Color Histogram, Cross Media Retrieval, Dictionary Learning, Minkowski Distance, And Term Frequency And Inverse Document Frequency.
Scope of the Article: Frequency Selective Surface