Description

Our goal was to detect the assigned object in images. The object had two modes, which were classified as ‘Car’ and ‘Standing.’ We planned to achieve object detection by using SIFT, which is a traditional feature detection method. SIFT was selected over methods such as SURF or ORB because we were interested in obtaining scale-invariant and accurate results, rather than faster results

My Role

• Conceptualized a simpler implementation of Pose bins (compared to the original algorithm)

• Generated Python scripts for implementing the Affine Parameters step of the SIFT algorithm