The images below show a basic example of using the program with an example image dataset.
First, load an image dataset. This can be done either by pressing the Open File or Open Folder
button or dropping a file or folder on the center part of the application. Images are either loaded
in grayscale or with colors. You can force to load a dataset in grayscale by clicking the
corresponding checkbox in the Loading section.
- If a folder is loaded, BLITZ checks all files inside and loads the ones with the most frequent suffix that appears.
- If an
.npyfile is loaded, the array can be of shape(N, m, n),(m, n)or(N, m, n, 3)for color images, whereNis the number of images andm, nis the shape (number of pixels) of the image. The checkbox grayscale influences the way an array with 3 dimensions is loaded. - If a video file gets loaded, BLITZ decides at which frequency to extract images, based on the given parameters 8bit, Subset ratio, Size ratio and Max. RAM.
Once a dataset is loaded, there are a number of metrics and information on the data directly visible:
The View operation set provides functions on changing the size or view of the image.
We call data manipulations along the time axis Reduction operations. This can be for example
computing the mean, maximum or minimum value of each single pixel accross all images.
Normalization subtracts or divides each image pixel by a certain value. This value often is chosen to be the mean of a certain subset of images.
If the actual size of an object in an image is known, the measuring tool can be used to measure distances in millimeters or angles in degrees.
For grayscale images, it is often useful to change the colortable to enhance visibility of low-value pixels (e.g. parts of the image that aren't enough lit-up).
