Keypoint-Matching / Movement delta: for a website app

Job ID: 31617233

Budget: $250 – $750 USD

I need an implementation which takes video data and detects/estimates the relative movement of the environment represented by the images. I reckon that this process involves key-point matching, so it seems realistic that it may involve OpenCV and WebAssembly (not 100% sure on this).

This project will be done in a GitHub repository. This repository contains some basic scripts as well as a page for the client. The "webmouse.ts" contains the WebmouseConnection class and the WebmouseVision interface. Since interfaces are not preserved from TS to JavaScript, the implementation should simply have a "getDelta" method.

The getDelta method accepts an image blob and returns a tuple-array containing the estimated dx and dy (pixels? percentage? we should discuss this) of the image compared to the previous image. Since this depends on a previous image, the returned [dx, dy] data of the very first call to this method should be [0, 0].

All of this will be done in a private (may go public later) repository. I will be designing the index page and all the other stuff which revolve around your code as well as the webmouse connections. All kudos/special-thanks for the heavy-lifting code will be given to you on the README, so please use your GitHub account to make the commits and changes (you will be invited as a collaborator).

Thank you for viewing
Related categories: JavaScript C++ Programming Typescript OpenCV