US 12,489,905 B2
Diffusion-based data compression
Noor Fathima Khanum Mohamed Ghouse, Amsterdam (NL); Jens Petersen, Amsterdam (NL); Tianlin Xu, London (GB); Guillaume Konrad Sautiere, Amsterdam (NL); and Auke Joris Wiggers, Amsterdam (NL)
Assigned to QUALCOMM INCORPORATED, San Diego, CA (US)
Filed by QUALCOMM Incorporated, San Diego, CA (US)
Filed on Aug. 29, 2023, as Appl. No. 18/458,006.
Claims priority of provisional application 63/410,581, filed on Sep. 27, 2022.
Prior Publication US 2024/0121398 A1, Apr. 11, 2024
Int. Cl. H04N 19/137 (2014.01); H04N 19/147 (2014.01); H04N 19/162 (2014.01)
CPC H04N 19/137 (2014.11) [H04N 19/147 (2014.11); H04N 19/162 (2014.11)] 30 Claims
OG exemplary drawing
 
1. An apparatus for processing image data, comprising:
at least one memory; and
at least one processor coupled to the at least one memory and configured to:
obtain a latent representation of an image;
process, using a decoder of a machine learning model, the latent representation of the image to generate an initial reconstructed image;
process, using a residual model, the initial reconstructed image and noise data to predict a plurality of predictions of a residual over a number of sampling steps, wherein the residual represents a difference between the image and the initial reconstructed image;
obtain, from the plurality of predictions of the residual, a final residual representing the difference between the image and the initial reconstructed image; and
combine the initial reconstructed image and the final residual to generate a final reconstructed image.