An optional feature of the i2k Retina API that is not available with the GUI version of i2k Retina is named Incremental Montage. A montage is typically created when a collection of images from a single study are passed to i2k Retina.

The processing time for creating a montage is influenced by the processing power of the computer, the quality of the images, the features visible in the images and the degree of overlap between images. Creating a blended image also adds time to the processing. When several images are supplied, it is not uncommon for a montage to take more than 20 seconds on a typical workstation.

By using the Incremental Montage library functions, an image capture application would asynchronously send images as they are captured. The library will return intermediate, or incremental, results that can be displayed to the retinal photographer who can visually confirm that sufficient images have been captured of the entire study area.

The time required to create each incremental result will also be a function of image quality, overlap and processing power. It is reasonable to expect an incremental result will be generated within two to five seconds.

After the final image has been added to the incremental montage function, a fully blended image of all the study images may be created.

The blended montage above illustrates the value of displaying an incremental montage to guide the retinal photographer. The images in this study were captured by a skilled and experienced retinal photographer, but without the benefit of any montage features.