18293678. BYSTANDER AND ATTACHED OBJECT REMOVAL simplified abstract (GOOGLE LLC)
Contents
BYSTANDER AND ATTACHED OBJECT REMOVAL
Organization Name
Inventor(s)
Orly Liba of Mountain View CA (US)
Pedro Velez of Mountain View CA (US)
Siyang Li of Mountain View CA (US)
Huizhong Chen of Mountain View CA (US)
Marcel Puyat of Mountain View CA (US)
Yanan Bao of Mountain View CA (US)
BYSTANDER AND ATTACHED OBJECT REMOVAL - A simplified explanation of the abstract
This abstract first appeared for US patent application 18293678 titled 'BYSTANDER AND ATTACHED OBJECT REMOVAL
Simplified Explanation: The patent application describes a media application that can detect a bystander in an initial image, generate a bystander box around the bystander, and segment the bystander and objects attached to them to inpaint the image.
Key Features and Innovation:
- Media application detects a bystander in an image
- Generates a bystander box around the bystander
- Creates localizer boxes around the bystander and attached objects
- Aggregates boxes to form an aggregated box
- Applies a segmenter to segment the bystander and objects
- Generates an inpainted image by replacing pixels within the bystander mask with background pixels
Potential Applications:
- Image editing software
- Surveillance systems
- Social media platforms for automatic tagging
Problems Solved:
- Efficiently segmenting bystanders and objects in images
- Enhancing image editing capabilities
- Automating tagging processes in social media
Benefits:
- Improved image editing accuracy
- Time-saving in tagging processes
- Enhanced surveillance capabilities
Commercial Applications: Automatic image editing software for social media platforms and surveillance systems.
Questions about Bystander Detection Technology 1. How does the media application differentiate between the bystander and other objects in the image? 2. What are the potential privacy concerns related to this technology?
Frequently Updated Research: Stay updated on advancements in image segmentation techniques and object detection algorithms to enhance the efficiency of the media application.
Original Abstract Submitted
A media application detects a bystander in an initial image. The media application generates a bystander box that includes the bystander, wherein all pixels for the bystander are within the bystander box. The media application generates localizer boxes that encompass the bystander and one or more objects that are attached to the bystander. The media application aggregates the bystander box and one or more of the localizer boxes to form an aggregated box. The media application applies a segmenter to the initial image, based on the aggregated box, to segment the bystander and the one or more objects from the initial image to generate a bystander mask, wherein the bystander mask includes a subset of pixels within the aggregated box. The media application generates an inpainted image that replaces all pixels within the bystander mask with pixels that match a background in the initial image.
