One danger that I don't think has been mentioned yet is seizure-inducing media. In the past, malicious users have been able to trigger seizures in users with epilepsy by uploading animated images designed to trigger them. This is a potential danger with any service that allows users to upload images, since even static images with certain dangerous patterns can potentially trigger seizures.
Vicki, once a manager of aircraft maintenance, is now charming Python & Ruby. She has coded OpenGates.dev, #VetsWhoCode bot, Code Questions bot, & LGBTQ of the Day Bot. They love relatable writing.
Location
Richmond, Virginia
Education
Self Taught Python/Ruby, BBA in Small Business Mgmt, AS in Production Mgmt
That's awesome! I've definitely thought about the hazards of some images, but I never expected there would be a project to detect them. I've starred it for future reference.
This is eye opening. I wonder how the algorithm for detecting these images would look? I did a quick search and couldn't see anything obvious in the results.
Vicki, once a manager of aircraft maintenance, is now charming Python & Ruby. She has coded OpenGates.dev, #VetsWhoCode bot, Code Questions bot, & LGBTQ of the Day Bot. They love relatable writing.
Location
Richmond, Virginia
Education
Self Taught Python/Ruby, BBA in Small Business Mgmt, AS in Production Mgmt
TensorFlow may imply neural networks, but only if you're aware of it. There are so many languages, frameworks, and libraries that you really can't assume that everyone would know what a certain language-specific library does. Even if it's a popular one, we can't know everything.
One danger that I don't think has been mentioned yet is seizure-inducing media. In the past, malicious users have been able to trigger seizures in users with epilepsy by uploading animated images designed to trigger them. This is a potential danger with any service that allows users to upload images, since even static images with certain dangerous patterns can potentially trigger seizures.
Detecting and removing seizure-inducing media could be one way backend devs can mitigate a major accessibility hazard outlined in WCAG 2.1 Guideline 2.3 – Seizures and Physical Reactions. For example, here's a Python project that aims to detect seizure-inducing media using TensorFlow.
That's awesome! I've definitely thought about the hazards of some images, but I never expected there would be a project to detect them. I've starred it for future reference.
This is eye opening. I wonder how the algorithm for detecting these images would look? I did a quick search and couldn't see anything obvious in the results.
TensorFlow implies neural networks, right?
TensorFlow may imply neural networks, but only if you're aware of it. There are so many languages, frameworks, and libraries that you really can't assume that everyone would know what a certain language-specific library does. Even if it's a popular one, we can't know everything.
Ah! Right, of course. Solid point.
That is a brilliant point! Thank you for sharing!!!
Woah I knew nothing about this, I'm going to do some research. Thank you very much Matt!
This community is so awesome 😭❤️