# Machine learning
Source: [https://docs.qualcomm.com/doc/80-70014-50/topic/machine-learning.html](https://docs.qualcomm.com/doc/80-70014-50/topic/machine-learning.html)
Qualcomm IM SDK provides support for end-to-end machine learning use cases that
include video pre-processing, model inference, output tensor post-processing, and inference
result overlay to stream.
The machine learning framework provides the following types of video analytics:
- Image classification
- Object detection
- Image segmentation
With machine learning plugins in the framework, use the following for inferencing:
- [TensorFlow Lite (TFLite)](https://www.tensorflow.org/lite/guide)
- [Qualcomm® Neural Processing SDK](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-2/overview.html) (Formerly
known as Qualcomm Snapdragon Neural Processing Engine (SNPE))
For each of the engines, delegates are enabled to accelerate model inference
performance.
## ML pipeline processing
The figure depicts
the pipeline in which the captured video/audio stream is processed among the
preprocessing, inference, and post-processing plugins.
Figure : ML pipeline
The table lists the sequential stages during which the video/audio stream is
captured, processed, and the output is either displayed on a screen or saved to a
file.
Table : ML pipeline processing stages
| Process | Description |
| --- | --- |
| Source |
- qtiqmmfsrc captures the YUV streams from
live camera source.
- The filesrc plugin captures any format from an offline video
that requires format conversion.
|
| Preprocessing | [qtimlvconverter](https://docs.qualcomm.com/doc/80-70014-50/topic/qtimlvconverter.html)does the following to prepare
the video stream for inferencing:
- Color conversion
- Resize
- Mean subtraction
|
| Inferencing |
- The preprocessed data is sent to the inference plugin in a
tensor format.
- The inference plugin sends this data to the machine learning
engines:
|
| Postprocessing | The output tensor gets passed down to postprocessing plugins: |
| Overlay | The overlay plugin that interprets the machine learning
metadata.
- qtioverlay draws appropriate overlay on
the buffer that allows the inference result to be visualized
in real-time.
- The stream can be rendered on to the display, streamed over
network, or encoded and stored to local storage.
|
| Sink |
- Waylandsink is used to display output on
a screen.
- Filesink is used to save the inferenced video to local
storage.
|
**Parent Topic:** [Qualcomm IM SDK GStreamer plugin architecture](https://docs.qualcomm.com/doc/80-70014-50/topic/architecture.html)
**Related Resources**
- [Machine learning use cases](https://docs.qualcomm.com/doc/80-70014-50/topic/machine-learning-use-cases.html)
Last Published: Oct 27, 2025
[Previous Topic
Graphics and display](https://docs.qualcomm.com/bundle/publicresource/80-70014-50/topics/graphics-and-display.md) [Next Topic
Base utility APIs](https://docs.qualcomm.com/bundle/publicresource/80-70014-50/topics/im-sdk-base-uitilities.md)