XYplorer 20.80 has been released on 10-Feb-2020. Here’s a quick introduction to the main new features:

• Duplicate Image Finder

• Duplicate Image Finder. Now you can find image dupes, and fast. Quite handy when sweeping your photo collection. You can even sort images by visual similarity, an interesting feature and hard to find elsewhere.

  XYplorer can generate an "image hash" (aka "perceptual image hash" or "fingerprint") for images by which you can find duplicate images, sort images by their visual similarity, measure the degree of similarity between two images (planned for a future version), and search for images that are similar to a given one. Unlike the common data hashes (MD5, SHA-256, ...) the image hash has an iconic relation to the visible image (the pixels), which means similar images have similar hashes.

  

  Two dupes found.

  
  

  When looking for duplicate images the setting of Tolerance comes into play. When Tolerance = 0 then all the pixels of the original full size images are compared and only perfect duplicates are listed. When Tolerance > 0 then the comparison is much more tolerant: it’s scale-invariant, brightness-invariant, contrast-invariant, saturation-invariant, and hue-invariant. The higher the value the larger the tolerance for other small differences in image details. The factory default is set to 9 which has turned out by experiment to be a good value yielding a minimum of false positives (found too many) and true negatives (missed out too many).

  Note that perfection is not realistic in this business but near-perfection is a reasonable goal. Even that will vastly speed up finding duplicate images, which by eye alone is near impossible if the number of images is higher than a few dozen.

  

  Dupe Finder Settings in the Info Panel (F12).

  
  

  Additionally there is a Special Property Column called "Image Hash" that can be used to sort images by similarity. An interesting feature and hard to find elsewhere.

  How to show the "Image Hash" column: To add the "Image Hash" column to the list you click menu command View | Columns | Add Column, then right-click the newly added "(Undefined)" column, then click "Select Special Property...", and select e.g. "Image Hash".

  In the column you can actually see the hash for each image:

  

  Sorted by similarity.

  
  

  Pro Tip: You can retrieve the hash of an image using the following script (run it through the address bar) while the image file is focused in the list:

  text "<prop #image.hash>";

  You then can search any location you like for similar images using the following search term in the Name field of the Find Files interface, or in Quick Search (F3). Note that 1b116dc5c3c5311b below is just an example for an image hash:

  prop:#image.hash:1b116dc5c3c5311b

  All images with the same hash will be found.