Samsung electronics co., ltd. (20240161455). APPARATUS AND METHOD WITH IMAGE MATCHING simplified abstract
Contents
- 1 APPARATUS AND METHOD WITH IMAGE MATCHING
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 APPARATUS AND METHOD WITH IMAGE MATCHING - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Original Abstract Submitted
APPARATUS AND METHOD WITH IMAGE MATCHING
Organization Name
Inventor(s)
Seung Wook Kim of Pohang-si (KR)
APPARATUS AND METHOD WITH IMAGE MATCHING - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240161455 titled 'APPARATUS AND METHOD WITH IMAGE MATCHING
Simplified Explanation
The abstract describes a method involving a convolutional network to extract feature maps and generate a matching point map between two images, refining it using attention between matching points, and extracting a matching point between the images from the refined map.
- The method involves extracting feature maps from multiple channels based on a convolutional network for two images.
- It generates a matching point map from the feature maps.
- The matching point map is refined using attention between matching points.
- Finally, a matching point between the two images is extracted from the refined matching point map.
Potential Applications
This technology could be applied in various fields such as image recognition, object tracking, and image comparison.
Problems Solved
This technology helps in accurately matching points between images, which can be challenging in tasks like image registration and object tracking.
Benefits
The method improves the accuracy and efficiency of matching points between images, leading to better performance in tasks like image recognition and object tracking.
Potential Commercial Applications
- Image recognition software
- Object tracking systems
Possible Prior Art
One possible prior art could be methods involving feature extraction and matching in image processing tasks.
What is the accuracy rate of the matching points extracted by this method?
The abstract does not provide specific information on the accuracy rate of the matching points extracted by this method.
How does the attention mechanism improve the refinement of the matching point map?
The abstract mentions using attention between matching points to refine the matching point map, but it does not detail how exactly this mechanism works to improve the refinement process.
Original Abstract Submitted
a processor-implemented method includes: extracting feature maps respectively corresponding to a plurality of channels based on a convolutional network with respect to two images; generating a matching point map from the feature maps; refining the matching point map by using attention between matching points comprised in the matching point map; and extracting a matching point between the two images from the refined matching point map.
