TO SAVE CUSTOM IMAGES: Create a new backdrop using only gray-scale colors, hit scan and then export! POST YOUR IMAGE DATA HERE: https://scratch.mit.edu/discuss/topic/119844/ SCAN IN HD: Click the up arrow next to scan for the option! ------------------------------------------------------------------------------------------- Hello! This is my second image scanner program. It uses an even more efficient scanning algorithm than before so that a whole 240x180 image can be scanned in under 20 seconds on a good computer. However, now I have included the ability to import and export image data into the program. Important: Although the ability to export and import images on Scratch has been approved by @paddle2see, you still upload any images at your own risk. Please report any posts with inappropriate images. The original image scanner can be found here: https://scratch.mit.edu/projects/63341952/ What makes the scanner so fast? 1) Updating the progress less often: Often, using pen or even stamp can greatly reduce the efficiency of a scanner. However, by only updating the progress at the end of every row, I am greatly reducing this factor. 2) Not checking for every color: There are 2 things going on here. Firstly, we exploit the fact that similar colors are marked as the same, so actually we don't need to increase the color value by 1 every time, and instead we can increase it by 8 which is much faster. Also, at the start, we do a fast scan of all the colors so that if there are any colors that don't turn up at all we don't need to test for them. 3) Checking against neighboring pixels first: Instead of having to go through the whole loop of colors, it is better to first start with checking with the color of the pixel to the left and the color of the pixel above. This is because often adjacent pixels have the same color. 4) Removing unnecessary functions: In the original scanner by @OxySean, time is wasted on every pixel when zeros are put in front of a number so that all the numbers look even. This is unnecessary and does not affect the result in any way. Credits: The scanner is based on 's original algorithm To get past the 10240 character limit, I am using a trick by The import text-box uses the technique demonstrated by here: The letter case detection script is also by ; the original project can be found here:
