18 lines
749 B
Markdown
18 lines
749 B
Markdown
|
# Azure Kinect Body Tracking OfflineProcessor Sample
|
||
|
|
|
||
|
|
## Introduction
|
||
|
|
|
||
|
|
The Azure Kinect Body Tracking OfflineProcessor sample demonstrates how to playback a recording Azure Kinect MKV file,
|
||
|
|
run through the body tracking SDK and store the body tracking results in a json file.
|
||
|
|
|
||
|
|
Notice: the current implementation is not the most efficient way to process a recording file offline. It simply
|
||
|
|
synchronously pushes and pops the capture to/from the tracker queue. To achieve the best performance, you would need to
|
||
|
|
create separate reader thread and process thread. And keep pushing new frames to the tracker without waiting for the
|
||
|
|
k4abt_pop_result to return successfully.
|
||
|
|
|
||
|
|
## Usage Info
|
||
|
|
|
||
|
|
```
|
||
|
|
offline_processor.exe <input_mkv_file> <output_json_file>
|
||
|
|
```
|