18379532. SEGMENTING OBJECTS BY REFINING SHAPE PRIORS simplified abstract (GOOGLE LLC)
Contents
SEGMENTING OBJECTS BY REFINING SHAPE PRIORS
Organization Name
Inventor(s)
Weicheng Kuo of Berkeley CA (US)
Anelia Angelova of Sunnyvale CA (US)
Tsung-Yi Lin of Sunnyvale CA (US)
SEGMENTING OBJECTS BY REFINING SHAPE PRIORS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18379532 titled 'SEGMENTING OBJECTS BY REFINING SHAPE PRIORS
Simplified Explanation
The abstract of this patent application describes a method for performing instance segmentation, which involves detecting and segmenting individual objects in an image. The method includes processing the image to identify a region depicting a specific object, obtaining example object segmentations, generating weight values for each example segmentation, determining a likelihood score for each pixel in the region using the example segmentations and weight values, and generating a segmentation of the object in the region based on the pixel scores.
- The method processes an image to identify a specific object region.
- Example object segmentations are obtained and weight values are generated for each segmentation.
- Likelihood scores are determined for each pixel in the region using the example segmentations and weight values.
- A segmentation of the object in the region is generated based on the pixel scores.
Potential applications of this technology:
- Object detection and segmentation in computer vision applications.
- Autonomous vehicles for identifying and segmenting objects in real-time.
- Medical imaging for identifying and segmenting specific structures or abnormalities.
- Robotics for object recognition and manipulation.
Problems solved by this technology:
- Accurate and efficient instance segmentation of objects in images.
- Overcoming challenges in identifying and segmenting individual objects in complex scenes.
- Reducing manual effort and time required for object segmentation tasks.
Benefits of this technology:
- Improved accuracy and precision in object segmentation.
- Faster and more efficient processing of large datasets.
- Automation of object segmentation tasks, reducing human effort and errors.
- Enhanced capabilities for various computer vision applications.
Original Abstract Submitted
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing instance segmentation by detecting and segmenting individual objects in an image. In one aspect, a method comprises: processing an image to generate data identifying a region of the image that depicts a particular object; obtaining data defining a plurality of example object segmentations; generating a respective weight value for each of the example object segmentations; for each of a plurality of pixels in the region of the image, determining a score characterizing a likelihood that the pixel is included in the particular object depicted in the region of the image using: (i) the example object segmentations, and (ii) the weight values for the example object segmentations; and generating a segmentation of the particular object depicted in the region of the image using the scores for the pixels in the region of the image.
