Chooch AI has formally launched the beta version of its Chooch IC2 mobile software development kit. The SDK is designed as an out-of-the-box AI solution that allows a company to train mobile applications to recognize specific items or people, drastically reducing the time and money it would take to develop such technology internally.
“Software developers can now integrate Visual AI into applications on mobile phones without piecing together a solution,” said Chooch co-founder and CEO Emrah Gultekin. “Before this, you needed to kluge together different deep learning components before you train a model. We’ve bundled these steps into a scalable solution, so you can send visual information into the prediction engine and use the output for your applications in real time.”
Those steps include labeling, data collection, and neural network selection, all of which are provided by Chooch’s complete end-to-end solution.
Facial recognition is one of the more obvious applications for the technology, but it can also be trained to identify everything from healthy cell tissue to ramen to machine parts. Once the AI knows what an image is supposed to look like, it can provide running metadata or notify the user whenever something seems off, giving it utility in industries ranging from surveillance to manufacturing to healthcare.
Chooch’s platform is one of several that demonstrates the growing appeal of customizable object recognition and machine vision, with Gopher Protocol announcing the development of a similar system back in August.