Cataloging images with Speech Labeling

Top  Previous  Next


With Speech Labeling, IDimager offers you a feature to catalog your images by simply speaking in the microphone of your computer. For Speech Recognition, IDimager uses the Windows Speech recognition software that comes with Windows XP, Vista, and Windows 7. There's no need to purchase any additional software to use it.


The concept behind it is simple. You select a few thumbnails and then pronounce the name of the catalog labels that you'd like to assign or revoke to/from your images.


There are several aspects to this which you need to consider. First of all you should train IDimager some basic commands that it uses. Once you've trained IDimager you can start using this feature. To better recognize catalog labels with speech commands it is also important that you train IDimager how you pronounce your labels.



The first time


The first time that you start the Speech Labeling feature, you are asked to start the training session. Activate the speech labeling feature by clicking the Speech button in the main toolbar.




Once the Speech Labeling feature is active, this button will stay toggled. Click it again to stop it again.


The first training session will now start:





There are 32 commands that you should train IDimager with and the progress can be followed by the number indicator in the title of Training dialog.


First a command is displayed and you're asked to pronounce that word.




If the word was correctly interpreted then a "Great!" indicator is displayed and after a short pause the next command in the training is automatically presented.




Repeat this for all 32 commands. And once you're done with each one of them the initial training session finishes.





The training session is a one time operation. However, you can always restart the training session in the future. Do this by starting the Speech Labeling feature and then pronounce the command "Start Training". That will restart the training session.


Start labeling


Now that you've trained IDimager for it should interpret your commands you can start using the Speech Labeling feature for the first time. You have already clicked the "Speech" icon in the main toolbar (see that it's toggled). Once the Speech Labeling feature is active you are ready to start labeling.


To do this, first select the thumbnails that you'd like to label and then pronounce the command:




That will open the labeling dialog:




We can now pronounce the name of the catalog labels that we'd like to assign.


When I pronounce my name in the microphone, it will automatically add the catalog label for it:




Repeat this for every catalog label that you'd like to have included:




When you pronounce a label name and there are multiple matches found in the database then you can easily pick the one that you need. A new dialog with green titles is opened with all found matches. Say the number of the one that you like and it will be picked and added to the existing list.






In this sample I have now collected three catalog labels in the Speech Labeling dialog.


It is possible to exclude labels by toggling the check boxes in front of each label name. This can be done by saying the number at its left. For instance, to toggle the second label I would pronounce the command TWO.



By pronouncing the command ASSIGN all of these checked labels will be assigned to the selected thumbnails.

By pronouncing the command UNASSIGN all of these checked labels will be revoked from the selected thumbnails.



Making a new thumbnail selection


Now that we've learned the basics on how to assign or revoke catalog labels for a selection of thumbs, it is good to know that you can also change the thumbnail selection while the Speech Labeling dialog is open. With the dialog open, you can say the command




And the selected thumbnails will be cleared.


You can navigate the focused thumbnail by using the commands







You can also select thumbnails with speech commands:







It is possible that the labeling dialog is overlapping the thumbnails which could block you from making a good selection. In that case use the speech command




to make the dialog transparent.




will then make it normal again.



(the faded dialog will make it possible to see the thumbnails in the back)