17525250. SYSTEM AND METHODS FOR MULTIPLE INSTANCE SEGMENTATION AND TRACKING simplified abstract (HUAWEI TECHNOLOGIES CO., LTD.)
Contents
SYSTEM AND METHODS FOR MULTIPLE INSTANCE SEGMENTATION AND TRACKING
Organization Name
Inventor(s)
Vishnu Sanjay Ramiya Srinivasan of Scarborough (CA)
SYSTEM AND METHODS FOR MULTIPLE INSTANCE SEGMENTATION AND TRACKING - A simplified explanation of the abstract
This abstract first appeared for US patent application 17525250 titled 'SYSTEM AND METHODS FOR MULTIPLE INSTANCE SEGMENTATION AND TRACKING
Simplified Explanation
The abstract of this patent application describes a method and system for multiple instance segmentation and tracking. Here is a simplified explanation of the abstract:
- The method involves sending an image to a backbone network to generate image feature outputs.
- These image feature outputs are then sent to a spatial attention module, which generates a feature map that identifies objects in the image.
- The feature map is also sent to a category feature module, which generates an instance category output indicating the objects present in the image.
- Additionally, the image feature outputs are sent to a mask generating module, which generates masks for the identified objects.
- The method further includes generating re-identification embedding information associated with the objects based on the image feature outputs.
Potential applications of this technology:
- Object detection and tracking in surveillance systems.
- Autonomous vehicles for identifying and tracking multiple objects on the road.
- Robotics for object recognition and manipulation in complex environments.
- Augmented reality applications for real-time object segmentation and tracking.
Problems solved by this technology:
- Efficiently identifying and tracking multiple objects in an image or video.
- Accurately generating masks for each identified object.
- Providing instance category output to indicate the type or category of each object.
- Generating re-identification embedding information for object tracking purposes.
Benefits of this technology:
- Improved accuracy and efficiency in multiple instance segmentation and tracking.
- Enhanced object recognition and tracking capabilities in various applications.
- Real-time processing of images or videos for immediate object identification and tracking.
- Potential for integration with existing systems and technologies.
Original Abstract Submitted
This disclosure provides for methods and a system for multiple instance segmentation and tracking. According to an aspect a method is provided. The method includes sending an image to a backbone network and generating image feature outputs. The method further includes sending the image feature outputs to a spatial attention module for generating a feature map associated with objects in the image. The method further includes sending the feature map to a category feature module for generating an instance category output indicating the objects. The method further includes sending the image feature outputs to a mask generating module for generating masks. The method further includes generating: the instance category output via the category feature module, and the masks via the mask generating module. In some embodiments, the method further includes generating re-identification embedding information associated with the objects based on image feature outputs.
