#فرقد_عدنان #farkadadnan #farkad_adnan
- facebook : https://www.facebook.com/profile.php?id=100002145048612
- instagram: https://www.instagram.com/farkadadnan/
- linkedin : https://www.linkedin.com/in/farkad-adnan-499972121/
The is a camera-based body movement tracking system that uses motion silhouettes to estimate the amount a person moves from a video of the person engaged in a task. A software based system for tracking movement over time
Authors: Farqad Adnan
Contact: farkad.hpfa95@gmail.com
This software was developed using Microsoft Visual C# 2010 on a Windows 10 64-bit machine.
System results:
The image in process is shown in the top left panel, while the corresponding motion silhouette is shown at the top right. The lower left panel shows an animated graph of the movement indicator over time. The program's controls, which allow the user to select the image to be processed, whether to show visualizations, and where to save the output files, are located in the lower right panel.
Sample output from the motion tracking algorithm.
Figure 
shows a sample output from the motion tracking algorithm. The panels on the left show single frames taken from the image sequence, and the panels on the right show the silhouettes of the corresponding motion. Pixels that have been shifted (that is, the places where movement occurred in the image frame) appear in white, while pixels that have not been shifted appear in black. In the bottom panel, there is a noticeable movement of the face and body. By contrast, in the middle panel, only some movement occurred, and in the upper panel, the body is still except for the eyes. As can be seen, the algorithm correctly filters background noise such as fixed pixels and light fluctuation, and detects both small movement such as eye blinks when the head is still (top panel), medium movement such as slight shifts in position (middle panel), and noticeable movements such as tilting Head and gestures (bottom panel).
On the left are single frames extracted from the image sequence, while the panels on the right display the silhouettes of the corresponding motion. Displaced pixels (that is, the places where movement occurred in the image frame) appear white; Pixels that are not shifted appear black.
Figure shows a time slice with 500 steps of the participant's time series. Video was recorded at 30 frames per second. The histogram displays movement in individual frames as a percentage of pixels per frame changing. The lower graph displays the estimated motion, which is the absolute difference in the ratio of pixels changing across adjacent frames. The periods of stable movement are reflected in the upper chart by small spikes in the lower chart. Small bumps indicate subtle changes in movement across adjacent tires. Large spikes in the lower graph indicate larger changes in movement across adjacent frames, reflecting sharp increases or decreases in the amount of movement. For example, in the upper chart there is a stable movement until approximately frame 30, at which point there is a sharp decrease in the amount of movement. This is reflected in the lower chart as a series of very small rallies during the period of stable movement, followed by a large rise when the movement suddenly decreases.
System requirements The program was tested on the 64-bit version of Windows 10. It should work on Windows 10 32-bit. It may work on older or newer Windows devices, but this has not been tested.
Microsoft .NET Framework 4.0. The program will prompt you to install this the first time you run it if necessary, and will take you to the [Microsoft .NET Download Page]
You may need to install [Visual C++ 2010 Redistributable Package] (https://www.microsoft.com/download/en/details.aspx?displayLang=en&id=5555)
You may need to install new video codecs so that the software can decode/encode video files. Installing a program like [ffdshow] can
Possible installation problems If you download the zip file, open the program, and run the demo, but the program crashes when trying to process the demo video file, you may not have proper video codecs installed. Try installing a program like [ffdshow], which installs several related codecs.
To select one AVI video file for processing, click "Select Input File". To select a directory for AVI video files for processing, click "Select input directory". To select a live cam feed to process, tap Capture From Camera.
You must specify a location to save the output files from the program. The program will automatically generate the motion time series and the midpoint of the motion time series. For example, if you choose to process a file called "video.avi", the program will create a file called "video-M.txt" with the motion time string, and a file called "video-C.txt" with the centroid time series. If you choose to save a video of animated silhouette frames, this file will be saved to "video-out.avi". If you choose to save the face region time series to a file, the output will be saved in 'video-ftxt'. If you are processing video from a live cam feed, you will have the option to choose if and where you want to save the recorded video.
Algorithm settings Motion history image duration The default value is 1000. The "memory" of the motion history image is in milliseconds, that is, the number of frames taken into account when determining the pixel density of the current motion history image. If the frame rate is 30 frames per second, setting this value to "1000" means that 30 frames will be used. bi-threshold BINARY_THRESHOLD: The default value is 30. After the silhouette mask is created from the latest video frame, the silhouette is selected to give a binary image. This value is the threshold. Haar Object Detection Settings Scale factor: default 1.1. How quickly the object detection scale increases with each pass over the image. Higher means faster, fewer passes, and you might miss things. Less means slower and more passes. The default OpenCv is 1.1 (10% increment on each pass). MIN NEIGHBORS: Default 2. Slash level to eliminate or keep groups of rectangles (found objects) based on the number of initial finds in the group. Combines groups greater than this value and discards groups with less value. If the detector misses many things.