FAQ

About AnTracks

Sample Applications

Documentation

Purchase & Licensing


"It's called 'AnTracks'. But can it also track other things than ants?"

"What are the system requirements to use AnTracks?"

"Installation worked fine, but when trying to launch AnTracks I get an error message. What can I do?"

"I want to run AnTracks on MacOS or Linux, is that possible?"

"Can I run multiple instances of AnTracks at the same time on one computer to process multiple videos at simultaneously?"

"I have a video file which AnTracks can't open. What's wrong?"

"Do I need image processing or computer vision knowledge to operate AnTracks?"

"I don't know how to set the parameters for my tracking applications. What can I do?"

"I want to track in infrared (IR) or ultraviolet (UV), can AnTracks handle that?"

"Should I label my animals with individual markers? / My animals are labeled with markers, can AnTracks recognize them?"

"How can I use the output data of AnTracks in other analysis software (like Matlab, IDL, Excel, etc)?"

"I have a video which suffers from camera shake and/or brightness fluctuations. Can I process that video in AnTracks?"

"Why can't I use JPG files as Exclusion Maps, StateMaps, etc?"

"Why does my histogram in the Event Analyzer have weird y-scaling?"

"I get 'bad frames' during processing. What are they, and how can I avoid them?"

"When using the Video Stabilizer, some frames seem distorted. What's wrong?"

"How can I improve the quality of my video recordings?"

"Is the source code for AnTracks available?"




"It's called 'AnTracks'. But can it also track other things than ants?"

Yes. 'AnTracks' actually stands for 'Animal Tracking Software' and can be used to track a large variety of objects, such as insects, fish, reptiles, mammals, cells in microscopy, humans, cars, etc. The type of objects you want to track does not matter for AnTracks. That being said, ants are indeed currently one of the most frequent applications for AnTracks, primarily because the technical challenges inherent to reliably tracking ants can not be met by many other tracking software.
[back to top]

"What are the system requirements to use AnTracks?"

AnTracks runs on Windows 7, Vista, XP. At least 2GB of RAM are recommended, and a minimum screen resolution of 1280x720 is required. AnTracks is not very CPU intensive, and most CPUs should work well. No special graphics hardware is required. Performance on Windows 8 is currently being tested.
[back to top]

"Installation worked fine, but when trying to launch AnTracks I get an error message. What can I do?"

If you get an error message like 'The application was unable to start correctly (0xc0150002). Click OK to close the application', it might be that your Windows version does not have the correct Visual Studio Redistributable Package installed. Make sure to install the proper 32-bit/64-bit version.
[back to top]

"I want to run AnTracks on Mac OS or Linux, is that possible?"

AnTracks natively runs on Microsoft Windows, and there are currently no versions of AnTracks for Mac OS or Linux. However, AnTracks runs with the Parallels Software for Mac or any other Virtual Machine on your Mac or Linux System.
[back to top]

"Can I run multiple instances of AnTracks at the same time on one computer to process multiple videos at simultaneously?"

Yes, it is possible to run multiple instances of AnTracks in parallel on the same machine. However, this is generally not recommended unless your computer has lots of memory (RAM), because memory usage can add up quite quickly when running more than one instance of AnTracks. Another limitation is the data input/output speed, which can become the bottleneck if multiple processes try to access a lot of data simultaneously. Also, note that it does not make sense to run more instances than your machine has CPUs.
In general, a much better option is to use the AnTracks Batch Processor to automate running multiple jobs sequentially.
[back to top]

"I have a video file which AnTracks can't open. What's wrong?"

AnTracks can read AVI files or image sequences as video input data. AVI files, however, are just a container format and the actual encoding format for the video (and audio) data is determined by the codec which was used to create the file. If an AVI file can't be opened by AnTracks, the respective codec could be missing on your system, or the codec might not be supported. Refer to the "Installation" section in the User's Manual for further information.
[back to top]

"Do I need image processing or computer vision knowledge to operate AnTracks?"

No special knowledge or skills in image processing or computer vision are required to use AnTracks. To get the best results for your detection/tracking application, some learning phase to get familiar with the software might be required, but the ramp-up phase is usually pretty quick and does not require any special technical knowledge.
[back to top]

"I don't know how to set the parameters for my tracking applications. What can I do?"

The tracking parameters have to be chosen appropriately for your individual application. Make sure you have taken the interactive Tutorial, and watched the tutorial videos on this website. See the User's Manual to get an idea what each PreProcessing Operation and parameter does and what its purpose is. Understand what each Frame Acquisition method does and try several ones for your application. Use the preview mode and the processing view to inspect what each operation and parameter does, then build your PreProcessing Operations queue accordingly.
[back to top]

"I want to track in infrared (IR) or ultraviolet (UV), can AnTracks handle that?"

Yes, that is possible with AnTracks.
More precisely, whether you can track in IR or UV depends on your video equipment - the actual video data (and thus the input to AnTracks) does not contain any information about the spectral range. Colors in a video can often be false colors, for instance when mapping infrared light to a dark red in the visual spectrum. Recording in IR or UV can be more difficult than recording in the visual spectrum, and have special hardware requirements.
[back to top]

"Should I label my animals with individual markers? / My animals are labeled with markers, can AnTracks recognize them?"

Using markers for tracking has advantages and disadvantages. Individual markers allow for a human to recognize and track individual specimen more easily, and for continuous observation over long time frames they can be the only way to keep track of individuals. However, in automated tracking, using markers is a trade-off. For the detection it often helps if the individual animals have a rather homogeneous appearance. For individual recognition, however, one needs rather distinct and heterogeneous appearance. Therefore, reliable tracking of individuals can actually work better without markers in some cases, whereas in other cases markers are essential. You have to be aware of this trade-off, and the question whether or not to use markers has to be decided case-by-case depending on your animals, experimental setup, and the particular observables you are interested in.
[back to top]

"How can I use the output data of AnTracks in other analysis software (like Matlab, IDL, Excel, etc)?"

AnTracks uses a generic text-based file format for Objects, Trajectories, and Events. These files can easily be imported into a program of choice. Alternatively, all data can be exported as XML files for further compatibility.
All images are in standard formats (PNG, BMP, JPG) and can simply be used with any image editing software.
For further information about file formats in AnTracks, see the respective section in the User's Manual.
[back to top]

"I have a video which suffers from camera shake and/or brightness fluctuations. Can I process that video in AnTracks?"

Yes. AnTracks has a Video Stabilizer module, which allows to correct for camera shake or movement. The quality of how good this works depends on the shake or motion in your video and on a few other things. Brightness fluctuations in a video can also be corrected by AnTracks, using the Brightness Equalizer module. See the respective section in the User's Manual for further information.
That being said, both these corrections introduce additional noise, and you will always achieve better results if your original video is already of good quality. For further information, see the section on "Noise" in the User's Manual.
[back to top]

"Why can't I use JPG files as Exclusion Maps, StateMaps, etc?"

JPG is a lossy compression format. This means that the pixel data is only an approximation of what the original image was. In particular, the pixel data already changes by just loading and saving a JPG file. With high quality settings, the differences are only very small and do not matter for most multi-media applications like photography. However, you can see artifacts of the lossy compression around sharp edges, for instance in the form of blurred and greasy fringes around text in a picture. Exclusion Maps and State Maps in AnTracks require exact pixel information, which JPG can not provide. The recommended file format for image files in AnTracks is PNG.
[back to top]

"Why does my histogram in the Event Analyzer have weird y-scaling?"

The histograms in the Event Analyzer are configurable by the user. You can control many parameters, including the number of bins (or the bin width), the range of the x- and y-axis, the frequency of the axislabels, the number of digits, etc. The following examples demonstrate good and bad choices for two parameters:

y-TickLabels and number of digits

The number of digits for the xTickLabels or the yTickLabels can be adjusted to show rounded or, depending on the range of the ticks, exact numbers. The example below demonstrates settings of 0, 1, and 2 digits for the case of 5 yTicks spaced equidistantly between 0 and 1.

Histogram with 2 digits for yTickLabels

The number of digits for the y-TickLabels are set to 2, displaying the exact values for these labels.


Histogram with 1 digit for yTickLabels

The number of digits for the y-TickLabels are set to 1, and the displayed values are rounded and not exact.


Histogram with 0 digits for yTickLabels

The number of digits for the y-TickLabels are set to 0. Thus, the displayed values are rounded towards integer numbers. For illustration, an extreme case is shown here where the rounded values are just 0's and 1's.


y-range and number of bins

An appropriate number of bins (or, alternatively, bin width) should be chosen depending on the data. Too few and large bins can obscure details in the data, whereas too many and small bins can make the "couting-aspect" of a histogram pointless. The example below shows the effect of setting the number of bins to 10, 20, 50, and 100, for the case of 23 data points. There is no generally "best" number of bins. Rather, an appropriate number of bins has to be chosen depending on the data and application in each individual case, as a compromise between robustness and details in the histogram.

Histogram with 10 bins

Histogram with 10 bins. A relatively low number of bins can obscure details, but has higher counts and shows less noise.


Histogram with 20 bins

Histogram with 20 bins. This histogram has a higher temporal resolution and shows more details than the histogram with 10 bins.


Histogram with 50 bins

Histogram with 50 bins. This histogram has an even higher temporal resolution, but each bin has a maximum of only 2 counts due to the small bin width. For most applications, a spacing like this would be too small.


Histogram with 100 bins

Histogram with 100 bins. In this extreme example, the histogram has so many small bins that each data point falls into a different bin. Thus, there are as many bars in the histogram as there are data points (23 in this example), and each bar has a height of exactly one count.



[back to top]

"I get 'bad frames' during processing. What are they, and how can I avoid them?"

The bad frame detection acts as a quality filter in AnTracks.
"bad frames" are frames that have unuseful or confusing data, and which therefore are ignored during processing. A video frame is considered a "bad frame" when certain filtering criteria are fulfilled. That is in particular if too many objects are detected in a frame, or if the objects are too large. Examples for situations when bad frames occur are when the camera's field of view was blocked by some object in front of it, or when (uncorrected) camera movement causes unexpectedly large movement of all objects.
Depending on your specific application, it is perfectly normal to have some bad frames. If you think that frames which should be okay are considered bad frames, then adjust the parameters in the "Object Detection" panel.
[back to top]

"When using the Video Stabilizer, some frames seem distorted. What's wrong?"

The Video Stabilizer module in AnTracks tries to correct for different types of motion - either purely translational motion, or affine transformations (translation, rotation, sheer, scaling). Depending on the geometry of your optical system, geometrical distortions (e.g. pin cushion or barrel distortions) will be present in the video. These distortions are static. Upon, e.g. translation of a video frame, these distortions can show up.
Also, distortions can be introduced with the Video Stabilizer if the camera exhibited motions that are not fully described by the motion model used for Video Stabilization. AnTracks performs global corrections based on local information (the anchor points). When the actual camera movement is a very complex mix of different types of motions, this can be insufficient.
To improve the performance of the Video Stabilizer, try choosing different anchor regions. Use anchors which are in different locations in the image, with a large separation between anchors. Try to change the size of the anchor regions, and that of the search radius. For further information, see the "Video Stabilizer" section in the User's Manual.
[back to top]

"How can I improve the quality of my video recordings?"

See On Video Recording for general tips on video recording.
[back to top]

"Is the source code for AnTracks available?"

No. AnTracks is a proprietary software product and the code is property of AnTracks Computer Vision Systems.
[back to top]