OS X Framework to parse Synopsis metadata, run spotlight searches, and sort results based on metadata contents, and optionaly analyze video and generate metadata dictionaries.
- OpenCV 3.3 + (included)
- zstd (included)
Optional:
- Tensorflow 1.2 + (Analysis only).
Synopsis framework can be compiled to provide just metadata reading / parsing / comparison - or in addition, providing analysis as well. These two build options are provided because Analysis requires the additional dependency on Tensorflow, which due to size and complexity needs to be compiled seperately.
To build a decode only Synopsis framework, simply check out the latest git repo, and compile the 'decoder' target. We include a small OpenCV2.framework pre-compiled in the git repo (without IPP)
To compile the with analyzer source, you need to also compile Tensorflow in a manner compatible with the Synopsis build process. You can optionally compile OpenCV with Intel Performance Primitives (IPP) for theoretical performance increases at a cost of 150MB additional binary size due to IPPICV library size. Our included OpenCV2.Framework has IPP disabled for easier deployment and smaller file size.
-
OpenCV - please follow the instructions on our wiki https://github.com/Synopsis/Synopsis-Framework/wiki/Building-OpenCV-for-Synopsis - and ensure if you code sign you enable other code signing flags --deep.
-
Tensorflow instructions coming.
Place the compiled OpenCV2.framework into the OpenCV subfolder so the XCode project finds it.