Negative values of MOTA, MOTP ... PeopleNet + KLT

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) RTX 2060
• DeepStream Version 5.1
• JetPack Version (valid for Jetson only)
• TensorRT Version 7.2.3.4
• NVIDIA GPU Driver Version (valid for GPU only) 460.56
• Issue Type( questions, new requirements, bugs) Question
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

I have evaluated People detection and tracking solution based on PeopleNet + KLT (nvtracker) using TrackEval on MOT Challenge in CLEAR metrics for data MOT20 training

Tracker output
MOT20-01.txt (737.8 KB) MOT20-02.txt (6.4 MB) MOT20-03.txt (4.5 MB) MOT20-05.txt (10.0 MB)

Tracker evaluation results
pedestrian_detailed.csv (1.2 KB) pedestrian_summary.txt (228 Bytes)

As can be seen in the attached. I don’t understand why i have sometimes ‘-’ or ‘+’ in MOTA There are also 0 values of MOTP, which signals an error in calculations (?)

2

I read this paper https://arxiv.org/pdf/2009.07736.pdf and found that MOTA is calculated as (TP-FP-IDS)/(TP+FN). So if the number of (FP+IDS) is greater than the number of TPs the score will be negative.

Now, if its possible could you kindly suggest how i can improve the results of MOTA and MOTP for PeopleNet + KLT for MOT20 training data

4

Hi rajkumarsaswamy,

Sorry for the late response, may I know if this is still an issue to support?

5

Hi @rajkumarsaswamy ,
Sorry for delay!
Discussed internally, we would recommend to try NvDCF tracker, it is better than KLT - more robust and has comparable performance.

Thanks!