An Efficient Analysis of Machine Learning Algorithms in CBIR
Palwinder Kaur, Rajesh K. Singh · 2020 International Conference on Computation, Automation and Knowledge Management (ICCAKM) · 2020
It is very challenging task to retrieve accurate image from huge digital images due to which much interest has been gain by CBIR. Feature based methods are prove to be more efficient as compared to text based query methods. As in case of CBIR a similarity between already stored database images and query by user can be extracted through shape, texture, color and structure like low level primitive features. This paper gives the idea about use of image retrieval in various fields and gives the problem related to Image retrieval. Its main problem is the retrieval of user requested images and search from database. A typical CBIR system automatically extracts shape, spatial, texture and colors information like visual attributes of every image in database. These extracted pixel values are stored in a different database within the system is called a feature database. There is increase in number of CBIR systems that result in increase of classification of these systems. One way of classifying these systems are texture, shape, colour and most frequently used features for feature extraction phase of CBIR systems. In last we review various works done by researchers in CBIR classification using Convolution neural network, Support vector machine, Random forest, Decision tree and K-nearest neighbour. This helps user in identifying the use of best classifier according to user need.