Hi there and welcome! This is Alex here. I am starting this personal/semi-professional blog on GitHub. Let us then launch this space with a banger! Today I would like to present you a project I have been working hard these past couple of months. OpenQVision is a Open Source project that allows Computer Vision to become more readable by non-programmer users, or even helps data scientists to check their neural network models on the fly.

OpenQVision Main Window

Currently, OpenQVision works with camera streaming, and you’ll need to set as input the path for your model files, i.e. weights, configuration and class names.

It is also possible to integrate it with external actions. The version uploaded triggers a Telegram Bot that send a capture after the detection runs for some user-defined amount of frames. Let us say, if a guitar is seen on the screen for more than 10 consecutive frames.

Telegram Bot Example

All right, before I forget, OpenQVision is written in C++. It is actually a port from its previous version, which was in Python language. For the sake of performance, I have managed to learn OpenCV in C++ and review my previous knowledge on Qt applications.

More details, access OpenQVision official repository https://github.com/dev-gazer/OpenQVision, message me on Twitter at @dev_gazer, find me on LinkedIn at https://www.linkedin.com/in/alexandremfaria/.

Cheers!

Now playing Slow Crush - Hush