Amazon Technologies, Inc.

Rohun Tripathi

Naveen Sudhakaran Nair

Angshuman Saha

Baris Gecer

Systems and methods are provided for automatic aspect ratio adjustment for video content. The system receives video content formatted in a first aspect ratio that is desired to be converted into a different aspect ratio. The video content is separated into segments, with each segment representing a different camera cut. A representative video frame from each segment is selected and fit to a canvas that is sized to the desired aspect ratio. The canvas is then provided to a generative AI model that is configured to 'de-noise' the pixels of the canvas located around the boundaries of the original video frame. This de-noising process modifies the pixels such that they more closely resemble a natural continuation of the original video frame. The final output is a modified video frame at the new aspect ratio that is less likely to provide the appearance of a modified video frame than conventional methods for adjusting the aspect ratio of video content.

Systems and methods for automatic aspect ratio adjustment for video content

Systems and techniques for validation and generation of localized content for audio and video are described herein. The systems and techniques provide for training of twin neural networks to evaluate performance characteristics, sometimes referred to as content-auxiliary characteristics, of a localized performance. The localized performance may be validated or improved by identifying misalignment in the performance characteristics to ensure that localized content preserves content as well as creative intent and performance ability in the final product. The machine learning models trained using the techniques described herein may be used in connection with auto-localization processes to automatically generate high quality localized audio and video content.

Growing community of inventors

Seattle, WA, United States of America

Rohun Tripathi