The Eye in AI

This 3D model is part of an originally developed activity for the Researchers Days, a science festival in Luxembourg, and currently used in our workshop on neural networks and Computer Vision called “Are computers blind?” (https://www.uni.lu/life-en/scienteens-lab/workshops/computer-science/)

The contents of the full printable:

-         Box composed of two trays and a slidable lid

-         9 white 1-tiles and 9 black 0-tiles

-         A 3x3 grid for the tiles and a “line” feature

The pdf document contains the instructions needed to do the activity with pupils/participants/visitors, showing step by step instruction of the activity.

To set everything up, remove the slidable lid and get out both the trays for 0’s and 1’s and fill them with the respective tiles, like this:

The tiles represent pixels in a binary image (i.e. pixels can only be turned either on (white = 1) or off (black = 0))

Get the 3x3 grid ready as the “screen” and provide the Haar kernel recognizing a vertical line.

To complete the activity, follow the steps provided on the instruction sheet. The first 3 steps help to familiarize the participant with the pixels and the screen grid. Step 4 is an exercise for the participant to follow themselves using the 3x3 grid and a part of a binary image.

Finally Step 5 goes one step further and exercises the calculations done on an image with greyscale pixels (where each pixel has a value between 0 and 255, representing the brightness of the pixel, where 0 means total dark and 255 means fully on, i.e. white).

If you have any questions or require further information regarding the background of the activity, please contact https://www.uni.lu/lcsb-en/people/christophe-stammet/ .