This paper proposes a new classifier named Extreme Learning Machine (ELM) on a hybrid framework for developing a Content Base Image Retrieval (CBIR) system to improve the accuracy problems faced with the earlier image retrieval system. This system mainly aims towards the accuracy with less consumption of time. In this system, Wang database is used with Local Binary Pattern (LBP), color moment, canny edge and region props for the extraction of texture, color, edge and shape feature respectively. After extracting all the features from the image, distance matrix will be determined to use it for further implementation. And then ELM classifier is used in this proposed CBIR to categorize all the images. Score Level Fusion is used as similarity measure for finding similar images. The obtained results proved that the accuracy and efficiency of CBIR system increased at a very high rate after using ELM classifier in terms of precision, recall, f-measure and retrieval time than just using similarity measure of the extraction features. The elapsed time and the average precision value is 0.277391 and 97.2500 respectively which is much accurate than the state-of-the-art techniques.

CBIR, color moment, canny edge detection, Local Binary Pattern(LBP), Extreme Learning Machine (ELM), Score Level Fusion

[1] L. K. Pavithra and T. S. Sharmila, “An efficient framework for image retrieval using color , texture and edge features R,” Comput. Electr. Eng., vol. 70, pp. 580–593, 2018.
[2] Mahajan, A. D., & Chaudhary, S. (2018). International Journal of Computer Sciences and Engineering Open Access Hybrid Features For Content Based Image Retrieval System, (10), 11–15.
[3] Samraj, J., & Dhivya, R. (2019). A Survey on Local and Global Feature Extraction Techniques in Content Based Medical Image Retrieval, (March), 251–265.
[4] Chavda, S. M., & Goyani, M. M. (2019). Survey Paper Recent evaluation on Content Based Image Retrieval, (4).
[5] Jeyakumar, M. N., & Samraj, J. (2019). Performance Analysis of Various Classifiers with Effective Dimensionality Reduction in Content-Based Image Retrieval, (March), 172–180.
[6] C. Paper and K. Kumar, “CBIR : Content Based Image Retrieval CBIR : Content Based Image Retrieval,” no. June, 2014.
[7] S. L. Dudhe, “‘ An ABIR And CBIR Fusion Based Techniques For Associated Image Retrieval ,’” pp. 3–7, 2016.
[8] M. Kaipravan, “A Novel CBIR System Based on Combination of Color Moment and Gabor Filter,” 1990.
[9] K. Kumar and J. Li, “COMPLEMENTARY FEATURE EXTRACTION APPROACH IN CBIR,” pp. 2–7.
[10] I. Conference, “CBIR by Cascading Features & SVM,” 2017.
[11] J. Pradhan, S. Kumar, A. K. Pal, and H. Banka, “CO CO,” Digit. Signal Process., vol. 1, pp. 1–24, 2018.
[12] C. Celik and H. Sakir, “Content based image retrieval with sparse representations and local feature descriptors : A comparative study,” Pattern Recognit., vol. 68, pp. 1–13, 2017.
[13] S. Mazharul, M. Banerjee, and S. Bhattacharyya, “Content-based image retrieval based on multiple extended fuzzy-rough framework,” Appl. Soft Comput. J., vol. 57, pp. 102–117, 2017.
[14] Deole, Pragati Ashok, and R. U. S. H. I. Longadge. "Content-based image retrieval using color feature extraction with KNN classification." IJCSMC 3.5 (2014): 1274-80.
[15] Katare, A., S. K. Mitra, and A. Banerjee. "Content-based Image Retrieval System for Multi Object Images Using Combined Features." 2007 International Conference on Computing: Theory and Applications (ICCTA`07).