Stands for IMage Detailer, a set of c programs to take large digital images and focus on a detail by creating a new image with a reduced version of the large image n the top left, and textual annotation on the top right and at the bottom the detail.

This can be done manually to a very high level of quality on a one-by-one basis quite judiciously with any number of programs that do digital imaging editing. But if the number of these images starts to increase, the task becomes quickly becomes time consuming to an exasperating level.

Enter imdetailer, takes on this task on a batch level, so that many hundreds and thousands of images can be detailed in one go. Quality if compromised without a doubt, but sometimes it may be at an acceptable. Also there is a manual element that can be introduced, by recording a text file specifying the point of interest or the centre of the detail.

Digital photos are everywhere now, people take very very many of them. Sometimes they have details that can be singled out, but these are often missed. So this repo is an effort to make the most of past digital photos.

aka. cairo_clip() followed by set_source, a negative positioning is noninuitively required. You can spend many hours working out why it doesn't work!

Photographs of coastlines from aeroplanes. It's very hard to identify where you are from the general picture. If you are able to focus on a detail, the issue is not solved, butg it can often be easier.

I was going to call it visual annotator or visual commentator not bad either, but I settled for image detailer or imdetailer because actually annotating is quite manual actually.

so the output images are going to be prefixed with "imda_" the a standing for automatic and "imdm_" for the idea being that the imda_ versions will use file name exif date and

But on which detail should be

The manual part is about clicking on an image and having the click positions recorded. Thi sis done via python's cv2 module. You need to have this installed on in your virtualenv. It will output the clieck positions into a text file with same base name as the image but with the "conft" extension added. Note: x and y values are put on separate lines! So, for two points you have four lines!

Rather than a field, this seems to be an instruction, and operates on the natural orientation of the image. It may be imposed by the camera, depending on the type of camera. Especially OpenCV (cv2) imshow and feh will obey this instruction, so cc1.py now picks it out and rotates (again) the image back.

However when using cv2 imshow, the window is taking on the size of the whole image, which blocks the current cursor position display which is on the bottom left. Neverthless, the click position goes well into the conft file.

TODO

• need to clean up cod
• produce a text based log file as it will be tedious going through all the _imd.jpg files.

iWell forget the text! In other words, it a picture tells a thousands words, why don't you use a subimage to do annotation. I'm already dong submimages BTW, so this is a cop-out in a way, but it does make sense. Cue inset0.c