openmht

Python module for multiple hypothesis tracking. Based on the article:

C. Kim, F. Li, A. Ciptadi and J. M. Rehg, "Multiple Hypothesis Tracking Revisited," 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, 2015, pp. 4696-4704. doi: 10.1109/ICCV.2015.533 URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7410890&isnumber=7410356

Note: This implementation utilizes motion scoring only (no appearance scoring)

Dependencies

Install the latest version of Python 3

Installation

$ pip install openmht

Input data

Format the input CSV columns with frame number and pixel positions using the examples under SampleData/ as a reference. The U,V values represent the 2D positions of objects/detections in that frame. A value of None in the output CSV indicates a missed detection. The Track column indicates the final track ID for a detection.

Parameters

Modify parameters by editing the file params.txt:

image_area: The image (frame) area in pixels (Default: 307200)

gating_area: Gating area for new detections (Default: 1000)

k: Gain or blending factor (Default: 0)

q: Kalman filter process variance (Default: 0.00001)

r: Estimate of measurement variance (Default: 0.01)

n: N-scan branch pruning parameter

Running:

OpenMHT takes 3 parameters: The input CSV, output CSV, and parameter file paths.

$ python -m openmht /Path/To/InputDetections.csv /Path/To/OutputDetections.csv /Path/To/ParameterFile.txt

Example output

Input

Output

Frame	U	V
0	-0.0411	0.208
0	9.97	10
0	15.2	14.9
1	14	13
1	-0.0099	0.00141
1	9.9	10.1
1	15.1	14.9
2	14.1	13
2	-0.009	0.00141
2	10	10.099
2	15	14.89

Frame	Track	U	V
0	0	-0.0411	0.208
0	1	9.97	10.0
0	2	15.2	14.9
0	3	None	None
1	0	-0.0099	0.00141
1	1	9.9	10.1
1	2	15.1	14.9
1	3	14.0	13.0
2	0	-0.009	0.00141
2	1	10.0	10.099
2	2	15.0	14.89
2	3	14.1	13.0