SIFT is an advanced image processing software designed to streamline analysis of high-speed images and videos for dynamic material characterization and failure experiments.


High-speed imaging technology can now provide high resolution images of material samples during a material characterization experiment up to and beyond failure. These images can then be analyzed for full-field strain analysis of the sample using various Digital Image Correlation (DIC) techniques. One of the pain points of working with high-speed images for DIC analysis is identifying the point of failure in a material test. This is a manual process which is cumbersome and time-consuming when working with many replicate image sets. SIFT is designed to automate identifying the point of failure using image processing techniques like dynamic feature tracking for failure point identification and to automate video generation from the still images.


  • Allows importing various image and video import formats.
  • Automatically finds the material sample in the image and identifies features to track within the sample.
  • Automatically identifies the point of failure by analyzing feature movement on a frame-by-frame basis.
  • Computes the pixel velocity chart that shows the user how the software determines the point of failure.
  • Exports videos to various video formats.


  • SIFT can autonomously identify the material within the frames by employing a thresholding technique to discern possible features of interest and meticulously removing outlier points. This culminates in the identification of a bounding box encompassing pertinent points of interest.
  • SIFT’s efficiency is established by its use of sparse optical flow analysis. This methodology, while delivering accurate results, also minimizes computational demands, enabling SIFT to track and analyze up to hundreds of points per second in a video stream in real time.
  • SIFT effectively stabilizes, tracks, and analyzes the points, transforming their positional data into velocity changes per frame. By systematically categorizing these points into groups based on their motion characteristics, SIFT can identify crucial peak points in the data, determining the most likely point of failure.