For the final project in the AI fundamentals course, I see the rubric provided although I’m not sure what the expectation is for complexity. As of right now, I’m planning to implement a project centered around basketball–would robust tracking of the basketball be sufficient, or should I aim for a more complicated feature such as shot detection or dribbling move recognition? I am certainly open to suggestions. Thank you!
Hi,
Suppose you are talking about object tracking on video.
You can find below a DNN-based object tracking tutorial:
<img src="https://github.com/dusty-nv/jetson-inference/raw/master/docs/images/deep-vision-header.jpg" width="100%">
<p align="right"><sup><a href="detectnet-tao.md">Back</a> | <a href="segnet-console-2.md">Next</a> | </sup><a href="../README.md#hello-ai-world"><sup>Contents</sup></a>
<br/>
<sup>Object Detection</sup></p>
# Object Tracking on Video
Although with the accuracy of modern detection DNNs you can essentially do "tracking by detection", some degree of temporal filtering can be beneficial to smooth over blips in the detections and temporary occlusions in the video. jetson-inference includes basic (but fast) multi-object tracking using frame-to-frame IOU (intersection-over-union) bounding box comparisons from [`High-Speed Tracking-by-Detection Without Using Image Information`](http://elvera.nue.tu-berlin.de/typo3/files/1517Bochinski2017.pdf) (DeepStream has more comprehensive tracking implementations available [from here](https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_plugin_gst-nvtracker.html)).
<a href="https://www.youtube.com/watch?v=L8vwuXKQrow" target="_blank"><img src=https://github.com/dusty-nv/jetson-inference/raw/master/docs/images/detectnet-tracking-pedestrians-youtube.jpg></a>
To enable tracking with detectnet/detectnet.py, run it with the `--tracking` flag.
``` bash
# Download test video
wget https://nvidia.box.com/shared/static/veuuimq6pwvd62p9fresqhrrmfqz0e2f.mp4 -O pedestrians.mp4
# C++
$ detectnet --model=peoplenet --tracking pedestrians.mp4 pedestrians_tracking.mp4
show original
Moreover, you can also find another sample with the multi-object algorithm:
When you observe something over a period of time, you can find trends or patterns that enable predictions. With predictions, you can, for example, proactively alert yourself to take appropriate action.
Thanks.
system
June 19, 2024, 5:53am
5
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.