Content based image retrieval (CBIR) relies on fetching relevant images from a dataset based on low level image features. Image features such as colour, texture and shape have been widely used in CBIR applications. These handcrafted features have been carefully designed and found to perform well in image retrieval task. The performance of these features greatly depends on the choice of the handcrafted features being used and domain knowledge. Hence there is a need to identify image features which are independent of domain knowledge and can be dynamically extracted from image data. Machine learning is a promising area here, since it focuses on learning representations from input data. Machine learning methods have been applied in various image processing tasks earlier. Convolution neural networks (CNN) models are able to create expressive features from image data and are successfully applied in image classification tasks. In this paper, we create a frame work to use CNNs to learn features from the image data and use these learned features for content based image retrieval. We test our proposed CBIR framework to retrieve images from a digital library database of art images. The results are compared against standard CBIR model which uses global colour histogram handcrafted feature. The results show that the learnt features extracted from a CNN model perform equally good as handcrafted features when applied to image retrieval task.

Content based image retrieval, Convolution neural networks, Machine learning, Global colour histogram

[1] Yong Rui and Thomas S. Huang, “Image Retrieval: Current Techniques, Promising Directions and Open Issues”, Journal of Visual Communication and Image Representation, 1999.
[2] Wei-Ying Ma and HongJiang Zhang, “Content-Based Image Indexing and Retrieval”, Handbook of Multimedia Computing, CRC Press, 1999.
[3] Lei, Z., Fuzong, L. and Bo, Z., “A CBIR method based on colour-spatial feature”, Proceedings of IEEE Region 10 Conference: TENCON 99, Multimedia Technology for Asia-Pacific Information Infrastructure, Vol 1: 166-169, 1999.
[4] Park, J., An, Y., Kang, G., Rasheed, W., Park, S. and Kwon, G., “Defining a New Feature Set for Content-Based Image Analysis Using Histogram Refinement “, International Journal of Imaging Systems and Technology - Multimedia Information Retrieval, Volume 18 Issue 2-3, 2008.
[5] N. Sharma, P. Rawa, and J. Singh, “Efficient CBIR using colour histogram processing,” Signal & Image Processing, vol. 2, no. 1, p. 94, 2011.
[6] H. Muller, N. Michoux, D. Bandon, A. Geissbuhler, "A review of content-based image retrieval systems in medical applications—clinical benefits and future directions", International Journal of Medical Informatics, 73, pp. 1-23, 2004.
[7] C.H. Lin, R.T. Chen and Y.K. Chan, “A smart content-based image retrieval system based on colour and texture feature”, Image and Vision Computing vol.27, pp.658–665, 2009.
[8] LeCun, Y., Boser, B., Denker, J.S., Henderson, D., Howard, R.E., Hubbard, W., Jackel, L.D.: Backpropagation applied to handwritten zip code recognition. Neural Comput. 1(4), 541–551, 1989.
[9] Bobić, V., Tadić, P., Kvaščev, G, “Hand gesture recognition using neural network based techniques.” in Neural Networks and Applications (NEUREL), 13th Symposium on (pp. 1-4). IEEE, 2016.
[10] Lawrence, S., Giles, C. L., Tsoi, A. C., & Back, A. D. “Face recognition: A convolutional neural-network approach.”IEEE transactions on neural networks, 8(1):98-113,1997.
[11] Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Rabinovich, A, “Going deeper with convolutions.” in Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9). 2015.
[12] S. Minaee, A. Abdolrashidi, and Y. Wang, "Iris recognition using scattering transform and textural features," in Signal Processing and Signal Processing Education Workshop (SP/SPE), 2015 IEEE, pp. 37-42, 2015.
[13] S. Minaee, A. Abdolrashidi, and Y. Wang, "Face Recognition Using Scattering Convolutional Network," arXiv preprint arXiv:1608.00059, 2016.
[14] Ivanova, Krassimira, “Content-Based Image Retrieval in Digital Libraries of Art Images Utilizing colour Semantics”, Institute of Mathematics and Informatics – BAS, Sofia, Bulgaria, 515-518. 10.1007/978-3-642-24469-8_62, 2011.
[15] D. Sherlin, D. Murugan, "A Case Study on Brain Tumor Segmentation Using Content based Imaging", International Journal of Scientific Research in Network Security and Communication, Vol.6, Issue.3, pp.1-5, 2018
[16] N. Rezazadeh, "Initialization of weights in deep belief neural network based on standard deviation of feature values in training data vectors", International Journal of Scientific Research in Computer Science and Engineering, Vol.5, Issue.4, pp.1-8, 2017