This directory contains simple examples and full applications that demonstrate how to use Scanner.
- Walkthrough.ipynb: an IPython notebook that goes through a simple application (shot detection) using Scanner.
- List of Tutorials: a set of well-commented files exploring different Scanner features in code.
If you want to run the notebook yourself so that you can interactively edit the code, run:
cd path/to/your/scanner/directory/
cd examples
jupyter notebook --allow-root --ip=0.0.0.0 --port=8888
Then in your browser, go to http://localhost:8888/notebooks/Walkthrough.ipynb and copy in the token from the console logs. Follow the instructions in the Jupyter notebook.
- face_detection: Detect all faces in a video, and then render a new video overlaying the face bounding boxes on the video.
- shot_detection: Segment a video into shots and then create a single image montage featuring one thumbnail for each shot. (Same application as the walkthrough.)
- hyperlapse: Create a stable timelapse video using the Hyperlapse algorithm.
- optical_flow: Use OpenCV to compute flow fields within a video.
- object_detection_tensorflow: Use Tensorflow's SSD Mobilenet DNN to detect objects.
- detectron: Use the Detectron object detection API for Caffe2 to detect objects.
- reverse_image_search: Search through a video to look for a query frame.
- depth_from_stereo: Compute a per-pixel depth image from two views of the same location.
- tensorflow: How to expose TensorFlow computations as Scanner graph operations.
- caffe: How to use Caffe nets as Scanner graph operations.
- python_kernel: How to implement Scanner graph ops in Python.
- halide: How to use Halide kernels as Scanner graph operations.