Bringing OpenCV to the Elixir World
summary
Evision generates NIFs from OpenCV headers to expose most OpenCV modules to Elixir and automates thousands of bindings instead of hand-writing them; it supports image I/O, feature stitching, SVM and SSD MobileNetV2 inference at ~18 ms on the demo machine and ~330–370 ms on a Raspberry Pi. It adds Evision generates NIFs from OpenCV headers and an Elixir backend with about 64 required callbacks, and integrates with Nx for zero-copy Tensor sharing so the library runs, is usable for a robot navigation showcase, and avoids a Python runtime.
AI-written summary · about
chapters (5)
- 0:0001:41 - Intro & Agenda
- 1:4206:00 - Evision. What is it and why it is Created?
- 6:0210:54 - Example of a NIF Library
- 10:5522:18 - Demo Time
- 22:1927:29 - Integration with Nx
related talks
26mEvision: An Evolving Elixir Library for Computer Vision Tasks
21mLeveraging Python for Real-Time Image processing
42mComputer Vision Using Go And OpenCV 3
30mIntroduction to Computer Vision with OpenCV
41mComputer Vision with GO and OpenCV
46mComputer Vision to Secure Your Surroundings with AI/ML Solution Built Using Open Source Tools at the Edge
04mEmbedded Executive LIVE
43mOh, Oh, Oh, It's Magic!