Wikipedia:Reference desk/Archives/Computing/2024 February 7
Computing desk | ||
---|---|---|
< February 6 | << Jan | February | Mar >> | Current desk > |
Welcome to the Wikipedia Computing Reference Desk Archives |
---|
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages. |
February 7
editFree AI-based app/algorithm to blur certain faces from an image?
editSpecifically, if the pictures in question depict children, toddlers, babies and whatnot. I work for city hall who has a Facebook page and for fairly obvious privacy reasons we wouldn't want to have the kids' privacy be violated or their likenesses unfairly exploited. Problem is that there's tons of images needed to be blurred and I don't have all the time in the world to edit them en masse. Is there something free and/or open-source to automate such tasks? Blake Gripling (talk) 06:23, 7 February 2024 (UTC)
- Are you looking for something that automatically identifies faces and blurs them, or something that merely does auto-blurring for each spot that you click? Nyttend (talk) 20:39, 7 February 2024 (UTC)
- The former; the latter's trivial enough for image editing suites such as FireAlpaca to do, but I am looking for an automated method to batch-blur photos based on the stuff I mentioned. Blake Gripling (talk) 23:53, 7 February 2024 (UTC)
- There are three main steps to a problem like this. I work with Python so I've found some packages which may work.
- Detect each face. You could use the face-recognition package. I haven't tested for accuracy but it has 51,000 stars on GitHub so it should be fine.
- Estimate the age of each face, and discard the faces of adults. I've found this package which is said to estimate ages, but I have never used it before and do not know the accuracy of it.
- Blur each underage face. To do this you can create a copy of the image and blur it. Then you can mask the two images using the bounds given by step one, which will leave only the faces blurred. I'd use Pillow for this.
- —Panamitsu (talk) 01:19, 9 February 2024 (UTC)
- Note: The package outlined in step 2 did not work for me so I just found a Caffee model online. —Panamitsu (talk) 03:04, 10 February 2024 (UTC)
- I'd be looking forward to some package that'll incorporate those algorithms into one (not-so) convenient UI. Blake Gripling (talk) 04:05, 10 February 2024 (UTC)
- Note: The package outlined in step 2 did not work for me so I just found a Caffee model online. —Panamitsu (talk) 03:04, 10 February 2024 (UTC)
- There are three main steps to a problem like this. I work with Python so I've found some packages which may work.
- The former; the latter's trivial enough for image editing suites such as FireAlpaca to do, but I am looking for an automated method to batch-blur photos based on the stuff I mentioned. Blake Gripling (talk) 23:53, 7 February 2024 (UTC)
🫶🏽 emoji
editWhat Unicode block includes the 🫶🏽 emoji? Per [1], it's a "Heart Hands", i.e. hands holding a heart symbol. Special:WhatLinksHere/🫶🏽 has nothing at all, which surprises me because most emojis are linked (at least) by an article on the Unicode block that includes them. Special:Search/🫶🏽 is mostly quotations of random people's Twitter posts. The exceptions, List of emojis and Apple Color Emoji, don't appear to mention this emoji at all (I just used Ctrl+F; apparently I've overlooking some hidden text) and don't provide Unicode-block information for any of the emojis that they list. Nyttend (talk) 20:44, 7 February 2024 (UTC)
- Symbols and Pictographs Extended-A. The code point is 1FAF6 (hex). --142.112.220.136 (talk) 22:22, 7 February 2024 (UTC)
- @Nyttend: I get 16 results on Special:Search/🫶🏽 and one of them is Symbols and Pictographs Extended-A. The hits on List of emojis and Apple Color Emoji are caused by 🏽. Those hits disappear for Special:Search/"🫶🏽". Help:Searching#Search string says: "Terms in the search string are subject to stem matching, except for anything included between double quotation marks." The result of stem matching can be hard to predict. PrimeHunter (talk) 15:35, 9 February 2024 (UTC)
Ctrl-Y timing problem
editIn Word, I've long used the Ctrl-Y function to avoid having to retype the previous command. It's always worked fine, but what's been happening a lot to me of late is that if I leave more than about 10 seconds between the original command and Ctrl-Y, or between one Ctrl-Y and the next, it doesn't work, and I have to retype my desired command and start again. Is this a feature of the latest version of Word, and if so, how do I extend, or turn off, the maximum time delay so that I can go away for, say, a few minutes and then return to my work and type Ctrl-Y with confidence that it will still repeat the previous command? -- Jack of Oz [pleasantries] 23:14, 7 February 2024 (UTC)
- I have not used Word in quite some time, so take this with a grain of salt, but my understanding is that ctrl-Y is a "redo" command. That is, it reverses the effect of the previous "undo" (ctrl-Z). If you haven't done a ctrl-Z but just want to repeat the previous command, I don't think ctrl-Y is the right command. I guess you could enter your command, then type ctrl-Z to undo it, then type ctrl-Y twice to restore the command and then repeat it. CodeTalker (talk) 01:26, 10 February 2024 (UTC)
- I highlight previous and hit` ctrl+x+v+v. Temerarius (talk) 15:23, 10 February 2024 (UTC)
- I hope you enjoy your salt, CodeTalker. From our article Control-Y:
- In most Windows applications this keyboard shortcut functions as Redo, reversing a previous Undo. In some programs such as Microsoft Office it repeats the previous action if it was something other than Undo. (my bolding of the relevant bit)
- It's worked for me precisely that way for at least 35 years. Now, something's changed. -- Jack of Oz [pleasantries] 08:34, 12 February 2024 (UTC)
- I hope you enjoy your salt, CodeTalker. From our article Control-Y: