daime - FAQ & bug reports A word on image editing ethics This message is quite long, but I feel it could be important especially for newcomers to image analysis and scientific visualization. Scientific research must be based on true observations and original (not inappropriately manipulated) data. Digital images such as fluorescence micrographs are raw scientific data, and any manipulation of these data is in principle questionable. Theoretically, image processing programs (including daime) could be misused for inappropriate image manipulation, which can lead to biased image analysis and visualization results. On the other hand, image editing steps such as noise reduction or background removal are often required for efficient image analysis and visualization. Removal of such features from scientific image analysis and visualization software would handicap the majority of users, whom I believe to be serious real scientists. But how do newcomers know which image editing steps are acceptable? Guidelines have been published, for example Rossner and Yamada, J. Cell Biol. 166: 11-15 (2004). In addition to these, I would like to give my personal view with respect to certain features of daime: 1. The original images (raw data) must be stored at a safe place, and all image processing should be performed on copies of the raw data only. Always store the original images (uncompressed Tiff files). The image file format of daime (STK files) is more convenient as it stores all segmentation and object classification data along with the images, but keep a backup of the original images in a general format that can be read by other programs, too. 2. Sometimes, the original images are not suitable for image segmentation. They may contain too much noise or the signal to noise ratio may simply be too low. daime contains a few functions for background and noise reduction. All these algorithms modify whole images and not only specific regions of an image. However, they may change features of the objects in an image such as brightness or edge sharpness. Therefore, all image analysis functions (except segmentation) should be applied on the original images and not on edited versions. In daime this is easy, because all segmentation data can be extracted from the edited and segmented images as object layer. This object layer can then be transferred back to the original images. Thus, the original images can be segmented and measured without changing them at all! 3. Micrographs often contain artefacts such as autofluorescent particles, which must be excluded from any measurements in order to obtain unbiased results. There are other situations when objects in the original images must be deliberately ignored - for example, one may be interested in measuring only single cells but not cell clumps. In daime, objects can be selected or rejected in the object editor. Rejected objects are never measured, but they are kept in the image for later reference. Thus, there is no need to permanently remove an object that should not be measured although daime offers such a feature (see topic 4). If the segmented images are stored in the file format of daime (STK), the current state of selection/rejection is kept for each object so that measurements can easily be repeated. 4. By using the object editor one can permanently delete individual objects or non-object pixels from images. There is also a function to eliminate pixels that fall into a user-defined intensity range. Why do such features exist in daime? (i) Because this is one way to eliminate noise (e.g. by removing very small objects, which cannot be cells, or non-object pixels, or very dark or bright noise). (ii) Because this may be required for the visualization of important features in 3D image stacks, which would otherwise be obscured by artefacts, other biomass, or parts of surfaces. A related feature is the clipping functionality of the Visualizer. Use these features with care and only when absolutely necessary. And provide a complete description of what has been done, for example in the figure legend in a publication. One could also publish the original images together with the modified versions to illustrate the changes and why they were needed. It would be unethical to delete objects that contradict a hypothesis, are there due to unspecific probe binding, or collide in any other respect with one's own expectations. The current version of daime displays a warning message when certain image manipulations are requested. 5. Like other 3D visualization programs, daime must perform calculations to create a 2D projection from a 3D dataset. It is important to understand that during this process, every single voxel gets an (artificial!) colour and opacity, which determine its appearance in the final image. The Visualizer allows the user to adjust many parameters, which affect the results of these calculations. Adjustments are often necessary to obtain a meaningful illustration of a dataset. There is nothing wrong with that as long as the result is still fact, not fiction. For example: rendering a semi-transparent or shaded 3D surface is a special way to depict an object and does not violate scientific rules as long as the real extent, shape, and position of the surface are shown. Again, a brief description (e.g., in a figure legend) may be the best way to avoid misunderstandings. I hope that these hints are useful and help to apply the features of daime efficiently and in accordance with scientific standards. Which daime version should I download? This depends on (i) whether your operating system is Linux or Windows, and (ii) which kind of installation you plan. If you use Windows, your best choice will most likely be the binary distribution of daime for Windows. It comes with a convenient setup program and should run out-of-the-box on all Windows 2000 and Windows XP systems (at least we hope so!). If you are an experienced Linux user or system administrator and want to install daime on a Linux machine, you will need to download and compile the source code of daime. Please note that this requires other software to be first installed on the Linux system. The installation manual explains the system requirements and the installation of daime for Linux. Do I need other imaging software to use daime? No. daime is a standalone, independent application and not just a macro for third-party commercial or freeware image analysis software. However, daime uses some libraries that provide basic functionality such as menus, dialog windows, and graphics output. Most users of daime will not need to obtain these libraries separately: they are already included in major Linux distributions, and we have included them for you in the installation package of daime for Windows. Depending on the Linux distribution, some parts must be installed from the Linux source CDs or DVD by using the package manager of the distribution. Please refer to the daime installation manual for details. Which operating system do I need to run daime? daime runs on commodity PCs and is currently available for Linux and for Windows 2000 or XP. Apple Macintosh is not supported yet (sorry Mac Users!). How is daime pronounced? daime is pronounced like "dymee". What is the price of daime? daime is free software published under the GNU Public License (GPL). The program and its source code can be downloaded free of charge. Users do not have to pay for the license to use daime. However, if you publish data or images obtained directly or indirectly by using daime then we ask you to cite the daime publication: Daims H, Lücker S, Wagner M. 2005. daime, a novel image analysis program for microbial ecology and biofilm research. Environ. Microbiol. 8: 200-213.  PDF
The GPL forbids incorporation of this software (or parts of it!) in commercial products or closed-source projects. The copyright holder is the author of the software. What is daime? daime is a digital image analysis and 3D visualization program optimized for processing images of microbial cells, which were stained by fluorescence techniques such as FISH with rRNA-targeted oligonucleotide probes. It is developed at the Department of Microbial Ecology, University of Vienna. Bug reports Questions, bug reports and comments can be sent to: 
|