Hi Jan, welcome to the MITK community!
> I have some challenges getting started. I have viewed the tutorials and > video, but that only gives a brief indication of what the software can do. Is > there a more starter-friendly, step-by-step guide / tutorial / youtube video > that a novice like can follow along? I have searched, but haven’t found any. What exactly are you looking for? If you're looking for more information on the segmentation/annotation plugin, you can press F1 when the tool is active. This will open a context specific help page right in the workbench. You can also click on Help > Open Help Perspective to get an overview of many more guides for different tools. To go back from there, click on Window > Open Perspective > Research. > The image sequence is named like 00000001.jpg to 00009999.jpg, they are shown > in proper sequence in the left panel, starting with the highest number at the > top and the lowest at the bottom, but for some reason the program selects > image nr. 25 (00000025.jpg) as the first image in the right panel to work on. > Why does it not start with the lowest number / file name? Why nr. 25? Is > there a setting to change this? The left panel is the Data Manager, which gives you an overview of all data that you have currently loaded or generated. The segmentation plugin is independent from the Data Manager selection. It has its own selection widget at the very top, were you need to select the image you want to work with. After you have selected the image you want to work with, you can create a segmentation for it by clicking on the small button right next to the segmentation widget below the image selection widget. Now you can start using the segmentation tools. I just noticed that the image list in the Segmentation plugin isn't always sorted like the Data Manager and we should synchronize that in the next MITK release. > Without a guide, I have tried some of the annotation tools. I can create a > contour with the lasso-tool, but as soon as I let the mouse button go, the > (green) contour disappears and nothing is left. This is different from what I > saw on the example video. I would expect a mask / fill to remain? Same > happens with the other tools. Is this because I loaded a B&W (jpeg) image? Is > it because it is a 2D and not 3D image / model? Imagine your images in the Data Manager like a stack of papers. When you are working with an image that isn't at the top, all the other images above it will block your view of the image of interest. In the Data Manager you can uncheck all the checkboxes of the images on top of your image of interest to make them invisible, or you can drag and drop your image of interest to the very top. That's why you didn't see your annotation. Also, you annotated something you saw on the top image instead of the image you were actually interested in. So you need to be careful here a bit. If you have the possibility to work with a 2D+t image (that is a single image file which contains all the timesteps of your sequence) instead of a sequence of several separate 2D images, I would recommend to go for that, as the complete image sequence would occur only once in the Data Manager and you can navigate with a slider through the time steps, which makes it much more comfortable. > the tutorial files are shown in a browser. It seems at the end of each, there > is no exit button to take the reader back to the main page from where he / > she started. The only way I could get out of the browser tutorial is by > pushing the back button as many times as there are pages in the tutorial. Is > that how it was intended, or did I miss something? Click on the MITK logo at the bottom left to go back to the Home screen. > Since the difference between individual frames is often small (limited > movement of the subject), I hope your software will allow automatic transfer > of annotations from one frame of the sequence to the next. This would > significantly reduce the manual labor. Can you confirm that this is possible? Unfortunately, there's not a copy button or similar, but you can work around by individually save a segmentation you have created (by right-clicking on it in the Data Manager > Save...) and then load it again. Now you have effectively copied your segmentation. In the segmentation view, you can associate that copy to your reference image of choice. You should consider renaming the image in the Data Manager to keep on overview. Best regards, Stefan ________________________________________ Von: Jan de Lange [janfdela...@gmail.com] Gesendet: Donnerstag, 4. August 2016 21:37 An: mitk-users@lists.sourceforge.net Betreff: [mitk-users] Annotations not visible Hello MITK team, Discovered your software today. Installed it (Win 7) and imported a sequence of 2D black and white video X-ray images of a spinal cord. My aim is to see if I can use your software to create a set of image annotations, i.e. image masks, as a training data set for machine learning. For this I need to manually segment video frames and save these (pixel / graphic files) together with the video frames data for supervised machine learning image recognition / segmentation. Since the difference between individual frames is often small (limited movement of the subject), I hope your software will allow automatic transfer of annotations from one frame of the sequence to the next. This would significantly reduce the manual labor. Can you confirm that this is possible? I have some challenges getting started. I have viewed the tutorials and video, but that only gives a brief indication of what the software can do. Is there a more starter-friendly, step-by-step guide / tutorial / youtube video that a novice like can follow along? I have searched, but haven’t found any. Without a guide, I have tried some of the annotation tools. I can create a contour with the lasso-tool, but as soon as I let the mouse button go, the (green) contour disappears and nothing is left. This is different from what I saw on the example video. I would expect a mask / fill to remain? Same happens with the other tools. Is this because I loaded a B&W (jpeg) image? Is it because it is a 2D and not 3D image / model? The image sequence is named like 00000001.jpg to 00009999.jpg, they are shown in proper sequence in the left panel, starting with the highest number at the top and the lowest at the bottom, but for some reason the program selects image nr. 25 (00000025.jpg) as the first image in the right panel to work on. Why does it not start with the lowest number / file name? Why nr. 25? Is there a setting to change this? One final note: the tutorial files are shown in a browser. It seems at the end of each, there is no exit button to take the reader back to the main page from where he / she started. The only way I could get out of the browser tutorial is by pushing the back button as many times as there are pages in the tutorial. Is that how it was intended, or did I miss something? I apologize for so many (trivial) questions, I hope you can help me out to get me started. Kind regards, Jan de Lange ------------------------------------------------------------------------------ _______________________________________________ mitk-users mailing list mitk-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mitk-users
