Image Detection Using Clustering and Scale Invariance

Image recognition is the process of comparing and identifying an object or a feature in a digital image or video. This concept is used in many applications, e.g., systems for factory automation, toll booth monitoring, and security surveillance. Color descriptors are employed to increase illumination invariance and discriminative power, but this technique when used in isolation does not lead to scale invariant image detection. SIFT, a feature detection mechanism, is scale invariant and is employed to improve search performance. The feature descriptors are clustered to derive a search index. When combined, these techniques provide improved matching to image queries