Halfway to allow the global annotation of images

Hi! My name is Neus. I am student of Telecommunications Engineering from the Telecommunication School (ETSETB) of the Technical University of Catalonia (UPC). Since summer 2006 I have been working as a software developer for the Department of Signal Theory and Communications (TSC) of the UPC, taking part in the i3media project and focusing my work in the development of GAT.

This summer I will work on the semantic part of GAT. On the one hand, I will allow the reading and writing of ontologies and annotations in OWL (Ontology Web Language) format. On the other hand, I will improve the navigation through the semantic classes tree. I will try to post every Thursday to report on the advances in this part of GAT.

But this week I have been giving support to Khristina. We are halfway to allow the global annotation of images! Below you can see a screenshot of GAT displaying a sequence of images selected.


Even though, this part it isn't ready yet. Today we have discussed with Xavi about some details. We have made a lot of important decisions.

First of all, we will remove checkboxes. Secondly, we will distinguish between terminal images (the first and the last images of the sequence) and intermediate images. We will mark the selection of terminal images with dark blue and the selection of intermediate images with clear blue. In addition, we have defined an active images that will be marked with another colour and that could be scroll by a left-click on an image or moving the wheel mouse.

There will be two forms to select a sequence:

1) Left-click on the first and the last images of the sequence. That will select the first and the last images as well as the intermediate images. If we make a new left-click on a non selected image, the active image will be scrolled, selecting the new intermediate images.
2) Left-click on the first image. That will select the image. Moving downwards or upwards the wheel mouse, the neighbour images will be selected or un-selected.

In both cases, the active image will be the last selected image. On the other hand, we will not allow the un-selection of intermediate images, only active image could be un-selected.

Finally, a right-click on the panel will validate the selection. The annotated images will be marked with another colour and they could not be added to a new annotation in order to avoid duplicates.

It is important to mention that in each annotation process the user could only select one sequence of images. That is, it will not be possible the selection of multiple sequences, the user will have to repeat the process for each sequence.

That is all we have decided today in the Skype-meeting (perhaps a flow chart would has better to show the whole annotated process… ). Well, the next step is to implement it. What a work!

More next week! :P