January 19, 2015 by Tomasz Malisiewicz

Today we’re glad to announce a fresh, spanking-new update to VMX, our flagship computer vision server. Today’s release focuses on usability and contains several new features to help you save time when training your own object detectors.

You can now search models inside your object library. This is useful when you have several hundred different detectors to choose from. You can sort by date, time, name, and size.

Image Uploading

You can now upload images into VMX via a drag-and-drop interface from your Desktop. Modern HTML5/js lets web-based applications feel like native applications, and uploading your own images into VMX gives you yet another way of training and/or testing your models.

Session Logging

You can monitor your object detection sessions remotely. We added a /session/#model-id/log.txt endpoint to our API which will return the last line of the internal log. This will report back what is being detected by the detector and useful for creating a visual log of what VMX sees.

Model Editor

Inside the model-editor, you can now remove positive and/or negative examples. Before you could only move examples between the positive and negative side, but now you can throw away any problematic “half-object” examples.

VMX Documentation

Documentation is now bundled with VMX. Just click the “?” icon inside VMX and it will bring you to the appropriate help page.

Download VMX today

If you haven’t yet purchased a copy, why wait? Getting a license today will give you access to today’s release, together with any future updates. (visit us on the web at https://vision.ai)

How to install/update on Mac OS X:

Download the new installer: VMX.pkg (~160MB)

If the installer detects a version of VMX already installed, it will save both your configuration file and your model library so you won’t have to activate again.

How to install/update on Linux and Windows:

If you are using Linux, you’ll need to first install Docker (or boot2docker on Windows).

If this is going to be your first time installing VMX, do the following:

git clone https://github.com/VISIONAI/vmx-docker-manager.git
cd vmx-docker-manager
./vmx start 3000

If you have a previous install of VMX, do the following:

cd vmx-docker-manager
git pull
./vmx update
./vmx start 3000

What’s next for vision.ai?

VMX API examples. Seeing VMX in action. DIY-home automation. We have been revamping our API and are currently building a small library of examples which show how to build simple Javascript applications using VMX.