US Patent Application 17716590. DATA AUGMENTATION FOR DOMAIN GENERALIZATION simplified abstract
Contents
DATA AUGMENTATION FOR DOMAIN GENERALIZATION
Organization Name
Inventor(s)
Laura Beggel of Stuttgart (DE)
Filipe J. Cabrita Condessa of Pittsburgh PA (US)
Robin Hutmacher of Renningen (DE)
Jeremy Kolter of Pittsburgh PA (US)
Nhung Thi Phuong Ngo of Karlsruhe (DE)
Fatemeh Sheikholeslami of Pittsburgh PA (US)
Devin T. Willmott of Pittsburgh PA (US)
DATA AUGMENTATION FOR DOMAIN GENERALIZATION - A simplified explanation of the abstract
- This abstract for appeared for US patent application number 17716590 Titled 'DATA AUGMENTATION FOR DOMAIN GENERALIZATION'
Simplified Explanation
This abstract describes a method and system for improving the performance of a machine learning model by generating training data. The process involves selecting a source image and a target image from a database. An image segmenter is used to create segmentation masks for both images, separating the foreground and background regions. The foreground and background regions are determined based on these masks. The target image's foreground is removed, and the source image's foreground is inserted into the target image, creating an augmented image with the source image's foreground and the target image's background. This augmented image is then added to the training data for the machine learning model.
Original Abstract Submitted
Methods and systems are disclosed for generating training data for a machine learning model for better performance of the model. A source image is selected from an image database, along with a target image. An image segmenter is utilized with the source image to generate a source image segmentation mask having a foreground region and a background region. The same is performed with the target image to generate a target image segmentation mask having a foreground region and a background region. Foregrounds and backgrounds of the source image and target image are determined based on the masks. The target image foreground is removed from the target image, and the source image foreground is inserted into the target image to create an augmented image having the source image foreground and the target image background. The training data for the machine learning model is updated to include this augmented image.
