Hi all, I have high speed videos (8 s, 240 fps) of mice with a tremor. The animal is held in a scruff and videotaped to isolate the tremor in the fore paw. I am new to signal processing and am wondering of ways to analyse the videos for getting amplitude and frequency of the tremor. Thanks for any help!
An interesting q. Suggest following:
1 Use imagej (FOSS)
2. Identify regions of interest (segmentation) from each frame.
3. Do correlation and plot to obtain freq and magnitude
Good Luck
Cheers
I need to see your data, but my gut feeling is:
- Subtract each frame from the previous one. Any signal left is caused by motion.
- Try to fixate the mice so that the tremor is in the same direction as the X or the Y axis of your images.
- If you did the previous preparation, then project the subtraction images on that axis, and glue the resulting 1D lines together, creating a new 2D image (see attached example).
- ... (more to do)
I suggest you read my paper about getting the phase and amplitude of breathing from chest X-ray images:
https://www.researchgate.net/publication/254861970_Extraction_of_the_Respiratory_Signal_from_Sequential_Thorax_Cone-Beam_X-Ray_Images
Conference Paper Extraction of the Respiratory Signal from Sequential Thorax ...
@ Lambert Zijp Good idea.!.Would you please post a readable and view able attachment. As it stands both are difficult. Could you add a little on so this can be tried out in a small example.
1."fixate the mice so that the tremor is in the same direction as the X or the Y axis of your images."
2."glue the resulting 1D lines together, creating a new 2D image "
Two problems need to be addressed 1. Noise 2. Registration(synchronization) of images with some landmark to exclude affine motion .
Thank you
Cheers
@Narasim: Press the download button; ResearchGate is not able to give a preview of a 'doc' file.
I don't think that noise will be a problem, but you might be right about movement of the whole mouse. I hope the 'scruff' (whatever that is) is tight...
Maybe put some unique color of paint on the paw, so that segmentation of the tremor is easy.
You are welcome Narasim.
Swati, feel free to ask questions, and maybe post an example AVI.
Please see if attached pdf which uses Scilab 5.3.3 (FOSS) is useful .It shows a pulsating circle and a projection onto Y axis as suggested by @ Lambert Zijp . It is just a toy example as I found Lambert Zijp 's method very appealing and was tempted to try it out.
Good Luck
Cheers
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication