PROJECT TITLE :

Enhancing Sketch-Based Image Retrieval by Re-Ranking and Relevance Feedback

ABSTRACT:

A sketch-primarily based image retrieval typically desires to optimize the tradeoff between efficiency and precision. Index structures are usually applied to massive-scale databases to comprehend efficient retrievals. However, the performance can be tormented by quantization errors. Moreover, the ambiguousness of user-provided examples could additionally degrade the performance, when put next with traditional image retrieval strategies. Sketch-based image retrieval systems that preserve the index structure are difficult. In this paper, we have a tendency to propose an efficient sketch-based image retrieval approach with re-ranking and relevance feedback schemes. Our approach makes full use of the semantics in question sketches and also the high ranked pictures of the initial results. We tend to conjointly apply relevance feedback to find a lot of relevant images for the input question sketch. The integration of the two schemes leads to mutual advantages and improves the performance of the sketch-primarily based image retrieval.