Jiatong Li

A new benchmark dataset and challenge results highlight the state-of-the-art in infrared image super-resolution for remote sensing, pushing the field towards practical applications.

A unified benchmark reveals the trade-offs between pixel-wise accuracy and perceptual realism in state-of-the-art image super-resolution techniques.

Generating realistic human-object interaction videos from text, images, audio, *and* pose is now possible, opening the door to automated content creation workflows.

Achieve SOTA detail-preserving human-product image generation by using a new attention mechanism and loss function, trained on a new large-scale dataset.