I think, a very common problem is to put particular points of two (or more) different pictures/drawables on top of each other. This could be done by a perspective transformation and is useful, if you for example want to create an animated gif out of many slightly different pictures (consider unwanted waggling of the camera) you took with your digicam. If you use the raw source pictures, there might be small "jumps" in the flow, but if you use a transformation to put particular points on top of each other, there is a smooth effect of motion.

I need this feature to create stereooscopic-images, using the "stereoscopic"-plugin. The output-image could be viewed using red/green glasses. With stereoscopic-images it is very important to adjust the two images (right, left) so that at least the points of the focused object are exact on top of each other.

Is there already a plugin which provides the needed feature? My recherche in the registry was negative, but may be I overlooked something.

Despite of the fact there might be a already existing plugin, I wrote a plugin which implements the following algorithm:
- You provide three source points in the left-eye-image/drawable which should be exact on top of the corresponding three points in the right-eye-destination-image/drawable.
- The plugin calculates a matrix for a affine transformation out of this data and performs the transformation.

Unfortunately, for the moment, the plugin is not very comfortable to use: you have to provide 6 points (x and y-value) in 12 spin-buttons and you have to enter all values by hand. Since I am not very firm with gtk and the gimp-tools, there is no way to enter these values just by clicking on the points in the displayed images. Could your help me in that point? I just need to know how to "grab" the point-coordinates of a click to an image! I want to take these coordinates and fill the spin-buttons with these data to provide more comfort. From the context of my plugin (the run-function), I don't know anything of the diplayed image (and the events that are activated in it), isn't it? All I know is the image_id and the drawable_id. Any hints are very welcome!!

My algorithm works very similar to the "perspective-transformation" in the toolbox. The main difference is that with perspective-transformation it is not possible to provide the source-points which are fixed to the corner points. You may only change the 4 destination points. Regardless, the user-interface of perspective-transform is very comfortable (grid-preview, move points in the picture) and I would be happy if my plugin could provide the same comfort.

Gimp-developer mailing list

Reply via email to