A Mechanism for Sketch Based Image Retrievals using Generalized Gamma Mixture Model (GGMM) and Relevance Feedback Mechanism

Abstract

The exponential growth of multimedia technologies facilitated the ease of developing various images having different shapes and scales. With the advent of mobile technologies, these messages so generated are being transmitted across the globe in different formats for different purposes. With these advancements methodologies thus developed for identifying or expressing subjects (individuals) their views by means of sketches. These sketch based images have many advantages, in particular, these images can be well considered in situations where the narration and capturing becomes difficult. The present article underlines a mechanism to interpret the images and also addresses the retrieval of such sketch based images using Generalized Gamma Mixture Model. The relevance feedback mechanism is utilized to retrieve more relevant to sketch based images based on the query image. The efficiency of the proposed word is evaluated using metrics like precision, recall, error rate, and retrieval accuracy.