Object Recognition and computer vision

3d Object Recognition using Point Cloud Library:

Objects are first scanned using a handmade rotating table and a KINECT.

Training the system for all the objects in the data-set (The data set used is for Amazon picking challenge 2015)
Testing the system on two objects with different orientations.

Master thesis:

Kassawat, Majd. "Application of some algorithms in the Point Cloud Library (PCL) for object recognition in robotic picking." (2016).

Object Detection using HAAR-like features

This method requires training a simple neural network using thousands of labeled images of the target object (Positive data-set) and thousands of background images (Negative data-set)

Auto blob detection using VISP

Tracking non-red ball with NAO

Color identification using NAO

Google Sites

Report abuse