This conversion is done by tesseract open source ocr. Has nothing to do with this except in the name.
I was working on an OpenGL extension which is a kind of video structure. It seems like it should work. I can paint an object in space and time. For instance I could define the video of a classroom with furniture and walls that have dimensions in time that span the entire video. Things that change would exist in various places at various times. In a way, that is what I was doing with my program. I specify what an object (in this case a genetic sequence) is, in time and space. In the case of a culture of cells, each cell would have a lifetime and position in the matrix. It would seem that using change in position would allow a higher density of data, at the expense of its dependency.
It seems to make sense , not as a graphics or video tool, but as a way to specify and use the matrix of n-dimensional space.