Image Registration

Model by Modzy

This model aligns one or more photos to a reference photo. It accepts JPEG, PNG, or TIFF encoded RGB images no larger than 5000 x 5000 pixels as input. It outputs the geometric transformation homography that transforms the ‘aligned’ image to the ‘reference’ image. Image registration can be used in many diverse applications. In biomedicine, it can be applied to find changes in scans and other images; it can also be used to detect changes or damage to buildings, vehicles, and other equipment.
  • Description

    Product Description


    0% Use Case Dependent – Further information here.

    Since this model is a preprocessing model for imagery, it does not have any associated and relevant performance. The model tends to perform better on images with scenes of flat grounds (e.g., fields, tops of buildings) instead of scenes with significant contour (e.g., hills, areas with many trees, riverbanks).


    This model was not trained on any dataset as it is a preprocessing tool that uses a transformation methodology build built into the OpenCV library. The algorithm first finds key-points and descriptors from each image using the ORB OpenCV function. It then matches the key features from both images using the brute-force matcher function in OpenCV (BFMatcher). Next, it finds the homography matrix and uses it to create a new image that represents the unaligned image being aligned to the reference image.


    This model is a deterministic computer vision algorithm implementation and has no associated training.


    Validation of the model output was done empirically, using several diverse test scenes.


    The input(s) to this model must adhere to the following specifications:

    Filename Maximum Size Accepted Format(s)
    reference 100M .tif, .jpg, .png
    align 100M .tif, .jpg, .png


    This model will output the following:

    Filename Maximum Size Format
    results.json 1M .json