Idea

Testing Graphical User Interfaces (GUI) is done very differently in commercial and academic environments. In commercial environments GUIs are tested with the simple and convenient Capture and Replay method. Academically developed GUI test methods use powerful but inconvenient model based approaches on the other hand.

My approach combines both methods to a convenient but powerful GUI testing method. The basic idea is the following:

1. The user executes example sequences on a GUI
2. A Model of the GUI is Generated
3. Test Sequences on the Model are Automatically generated
4. The Test Sequences are automatically executed on the GUI

The GUI Models stay compact and clear because they use hierarchies and orthogonality.

A systematic combinatorial approach generates the Test Sequences. This keeps the sequences complete but lean.

Features

Convenient but Powerful GUI-Testing • Capture-&-Replay Approach • Model Based • Systematic Test Sequence Generation • Automatic Model Generation • User Friendly GUI Models • Complete but Compact Test-Sequences • Test Arbitrary GUIs

Prototype

As part of my final thesis I also developed a prototype. The prototype is a Plugin for the commercial testing tool TESTONA. You can see it in action here:

[Archived: embedded content unavailable]

Technologies & Methodologies

Context

This was a self organized project as an employee in the Berner & Mattner enterprise.

To realize this project I worked together with many other teams. I modified and integrated parts of their work into my GUI Testing Tool.

Details

I released a scientific paper   as part of this project. I also released the final thesis  and the final presentation  (in german).

[Archived: embedded content unavailable]

The post GUI Testing – A Concept and Tool to Test any Graphical User Interface appeared first on Oliver Stadie - Portfolio.

Context

One Idea of Augmented Reality is the mirror below. This mirror shows an altered version of you. Thus it allows you to select your clothing in a menu. You don’t need to change it in reality.

One component of this augmented mirror is an algorithm to analyse the real clothing.

I developed an algorithm which analyses the structure of clothing and fabric to find patterns.

Idea

The algorithm takes an image of the clothing as input and extracts the structure (mesh) and texture seen in the image.

Extract the Structure (Mesh) of  an Image

Extract the Texture of an Image

Using this structure new textures could be applied to the structure to complete the mirror.

Details

I published the complete student research paper  (in german).

The post Computer Vision – Analyzing the Structure of Clothes appeared first on Oliver Stadie - Portfolio.

Background

The Idea was to merge the information of multiple sensors (acceleration, pressure, camera, etc.) instead of using every sensor on it’s own, to improve the grasping and walking abilities of humanoid robots. Also additional sensors where added.

Nao Robot with additional sensors on his arms

Technologies

• Hardware: Nao Robots, additional modular sensors (e.g. acceleration)
• Software: C++

Results

The research resulted in promising new approaches for adaptive walking an grasping. A detailed paper on one of the grasping approaches was released: 

Nao Robot grasping objects of different weight and stability using the same adaptive algorithm and additional sensors.

The post AI & Robotics appeared first on Oliver Stadie - Portfolio.