lukevanin / SkyLight

Performant computer vision for Apple devices using Swift and Metal

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

SkyLight Computer Vision

Copyright © 2022 Luke Van In

SkyLight is a testbed fo computer vision and AI for Apple devices (iPhone, iPad, Mac). It consists of computer vision algorithms, such as feature detection and optical flow estimation, using Metal compute shaders that run on GPU hardware, with the aim of providing interactive and real-time performance on consumer mobile hardware (e.g. iPhoneX / iPad).

SkyLight is currently actively undergoing research and development. Collaboration is highly encouraged. Bug reports, feature requests, and pull requests are welcome.

If you use this code in research or in a commercial or open source project, I would love to know. Please get in touch in GitHub or on Mastodon or LinkedIn.

LICENSE:

By downloading or viewing this collection of work you agree to the terms of this license.

This license must accompany all work and research derived from this collection of work.

This license may be amended or revoked at any time without notice.

This work is provided as-is without any warranty.

This work may not be redistributed.

Any works which make use of this work in part or in whole must make their source code publically available.

Permission is granted for personal and academic use of the "Approximate Nearest Neighbor Trie" data structure and search algorithm. The algorithm may not be used for any commercial undertaking without prior consent from the author.

Works and publications that describe this work and derivations of this work must cite this repository and the original authors.

CITE:

"Approximate K Nearest Neighbors, Luke Van In, 2023"

About

Performant computer vision for Apple devices using Swift and Metal


Languages

Language:Swift 98.1%Language:Metal 1.6%Language:Objective-C 0.2%Language:C 0.1%