LG’s new neural network can create text descriptions from images
[ad_1]
Korean company LG at the Computer Vision and Pattern Recognition 2023 conference in Canada introduced Captioning AI neural network, which can create text descriptions for pictures. The developers note that AI can help organize photo databases.
The company considers its product unique, as it can even describe what it has never seen, for example, identify a landscape or a person from a photograph. AI can generate text descriptions for over 10,000 images in less than two days. The system is powered by LG AI Research’s Zero-Shot Image Captioning technology, which allows a neural network to describe scenes or objects based on previous experience. Also, AI can describe the background, the activity of the characters and the interaction of elements.
[ad_2]
Source link
