New AI Model Can 'Cut Out' Any Object Within an Image (arstechnica.com) 19
Meta has announced an AI model called the Segment Anything Model (SAM) that can identify individual objects in images and videos, even those not encountered during training. From a report: According to a blog post from Meta, SAM is an image segmentation model that can respond to text prompts or user clicks to isolate specific objects within an image. Image segmentation is a process in computer vision that involves dividing an image into multiple segments or regions, each representing a specific object or area of interest. The purpose of image segmentation is to make an image easier to analyze or process. Meta also sees the technology as being useful for understanding webpage content, augmented reality applications, image editing, and aiding scientific study by automatically localizing animals or objects to track on video.
Typically, Meta says, creating an accurate segmentation model "requires highly specialized work by technical experts with access to AI training infrastructure and large volumes of carefully annotated in-domain data." By creating SAM, Meta hopes to "democratize" this process by reducing the need for specialized training and expertise, which it hopes will foster further research into computer vision. In addition to SAM, Meta has assembled a dataset it calls "SA-1B" that includes 11 million images licensed from "a large photo company" and 1.1 billion segmentation masks produced by its segmentation model. Meta will make SAM and its dataset available for research purposes under an Apache 2.0 license. Currently, the code (without the weights) is available on GitHub, and Meta has created a free interactive demo of its segmentation technology.
Typically, Meta says, creating an accurate segmentation model "requires highly specialized work by technical experts with access to AI training infrastructure and large volumes of carefully annotated in-domain data." By creating SAM, Meta hopes to "democratize" this process by reducing the need for specialized training and expertise, which it hopes will foster further research into computer vision. In addition to SAM, Meta has assembled a dataset it calls "SA-1B" that includes 11 million images licensed from "a large photo company" and 1.1 billion segmentation masks produced by its segmentation model. Meta will make SAM and its dataset available for research purposes under an Apache 2.0 license. Currently, the code (without the weights) is available on GitHub, and Meta has created a free interactive demo of its segmentation technology.
Prior art (Score:5, Funny)
Here [gazetadopovo.com.br].
Couldn't come up with an 'In Soviet Russia' quip in time to make first post.
Re: (Score:1)
It's just in time for political pogrom season!
Re: (Score:2)
Coolbeans (Score:3)
Wrong approach (Score:5, Interesting)
That way I can cut out all memory of my deadbeat biological father.
Or to remove people's memories of me doing cringy shit as a kid.
Removing memories is the wrong approach.
What you want to do is to integrate your experiences of a deadbeat father into your own personality to make it better. Go through all the experiences, make a list of general observations (overall rules that predict his behaviour and your situation, based on statistical evidence from the experiences), assess each one and then decide whether you will be like that as an adult, or if different, how you will act differently as an adult.
Your life is now up to you to construct and manage - you can't blame things on your parents (or anyone else in your past) forever.
(It helps to actually do this, and to write it out longhand (ie - not typed into a computer). There are ways you can structure the exercise methodically (find them online) such as "divide your life into 7 phases and write a couple of paragraphs describing each phase" that can help you get a handle on what to do and how to do this.)
(People who take the time to do this are statistically more likely to have success in life. College students who take the full course of this, on average, go up 1 grade level in achievement.)
Re:It's called an edge detector (Score:4, Insightful)
in soviet Russia we cut you out and don't pay no a (Score:1)
in soviet Russia we cut you out and don't pay no abode fees for our Photoshop
Great for capchas! (Score:2)
Now, when I can't seem to solve a captcha, I can just use AI to solve it for me!
Re: (Score:2)
What do you think answering all of those captcha images was being used for? That's the training set.
Masking (Score:2)
What's the cost per image? (Score:1)
Saw a picture today (Score:2)
What about inserting objects? (Score:2)
Don't accuse me of having a dirty mind!
The github repo (Score:2)
Here it is: https://github.com/facebookres... [github.com]