SIFT as a Service: Turning a Computer Vision algorithm into a World Wide Web Service Problem Statement SIFT as a Service The Scale Invariant Feature Transform.

SIFT as a Service: Turning a Computer Vision algorithm into a World Wide Web Service

Problem StatementThe Scale Invariant Feature Transform (SIFT) is probably the most popular and widely applied feature detectors which assists a variety of applications including object recognition, image registration, object localization, image forgery detection, and 3D surface reconstruction. SIFT as a Service is a reusable, platform independent, and highly available software component which can aim Rapid Application Development (RAD) and fast prototyping for computer vision students and researchers all around the world. An Internet connection is all they need!

Ahmad Pahlavan Tafti and Zeyun Yu Department of Computer Science, University of Wisconsin Milwaukee, WI, USA.

Objectives·To facilitate Rapid Application Development and fast prototyping for computer vision researchers. · To provide application-to-application interaction for a highly demanded computer vision application. · To make the SIFT algorithm available through both human-oriented and application-oriented interfaces.

Approach

Figure 2. SIFT as a Service: Service architecture.

Figure 1. SIFT as a Service: Abstract view. Different group of users can get into the system using a browser, and developers may build a consumer application to call the SIFT as a Service without employing a browser.

Experimental Validations

Figure 3. Accuracy in Feature Points Detection. (A) Results obtained by the original SIFT (#features: 1638). (B) Results obtained by the SIFT as a Service (#features: 1627).

Figure 4. Accuracy in points matching. (A) Results obtained by the original SIFT software (#matches: 166). (B) Results obtained by the SIFT as a Service (#matches:158).

Impact

Figure 5. Impact of the SIFT as a Service. Number of users of the SIFT as a Service since December 2014.