View source for International business machines corporation (20240135675). DETECTING FINE-GRAINED SIMILARITY IN IMAGES simplified abstract

=DETECTING FINE-GRAINED SIMILARITY IN IMAGES=


==Organization Name==

[[:Category:international business machines corporation|international business machines corporation]]


[[Category:international business machines corporation]]

==Inventor(s)==

[[:Category:Fei Wang of Dalian (CN)|Fei Wang of Dalian (CN)]][[Category:Fei Wang of Dalian (CN)]]

[[:Category:Xue Ping Liu of Beijing (CN)|Xue Ping Liu of Beijing (CN)]][[Category:Xue Ping Liu of Beijing (CN)]]

[[:Category:Dan Zhang of Beijing (CN)|Dan Zhang of Beijing (CN)]][[Category:Dan Zhang of Beijing (CN)]]

[[:Category:Yun Jing Zhao of Beijing (CN)|Yun Jing Zhao of Beijing (CN)]][[Category:Yun Jing Zhao of Beijing (CN)]]

[[:Category:Kun Yan Yin of Ningbo (CN)|Kun Yan Yin of Ningbo (CN)]][[Category:Kun Yan Yin of Ningbo (CN)]]

[[:Category:Zhi Xing Peng of Beijing (CN)|Zhi Xing Peng of Beijing (CN)]][[Category:Zhi Xing Peng of Beijing (CN)]]

[[:Category:Jian Long Sun of Langfang, Hebei (CN)|Jian Long Sun of Langfang, Hebei (CN)]][[Category:Jian Long Sun of Langfang, Hebei (CN)]]

==DETECTING FINE-GRAINED SIMILARITY IN IMAGES - A simplified explanation of the abstract==


This abstract first appeared for US patent application 20240135675 titled 'DETECTING FINE-GRAINED SIMILARITY IN IMAGES

==Simplified Explanation==

The patent application describes a method for detecting fine-grained similarity in images by analyzing key areas and generating feature descriptors for comparison.

* Core area of search image determined by generating image salient map
* Feature descriptors generated from core area of search image
* Capsule vectors generated from keypoints of feature descriptors
* Comparison of capsule vectors between search image and dataset images
* Selection of images with fine-grained similarity based on bundled similarity score

== Potential Applications ==
This technology could be applied in image search engines, content-based image retrieval systems, and image recognition software.

== Problems Solved ==
This technology helps in identifying images with subtle similarities, which can be challenging for traditional image recognition algorithms.

== Benefits ==
The method allows for more accurate and efficient image matching, enabling better search results and content organization.

== Potential Commercial Applications ==
Commercial applications include visual search platforms, e-commerce product recommendations, and security systems for image verification.

== Possible Prior Art ==
One possible prior art could be traditional image matching algorithms that rely on pixel-level comparisons rather than feature descriptors and capsule vectors.

=== What are the limitations of the proposed method in detecting fine-grained similarity in images? ===

The proposed method may have limitations in handling large datasets due to the computational complexity of generating feature descriptors and comparing capsule vectors.

=== How does this technology compare to existing image recognition systems in terms of accuracy and efficiency? ===

This technology offers improved accuracy in detecting fine-grained similarities but may require more computational resources compared to traditional image recognition systems.


==Original Abstract Submitted==

detecting fine-grained similarity in image includes determining a core area of a search image by generating an image salient map from a plurality of layers of the search image and determining a connected area based on the image salient map. feature descriptors are generated from the core area of the search image. a plurality of capsule vectors are generated from different ones of a plurality of keypoints of the feature descriptors. capsule vectors of the search image are compared with capsule vectors of each image of the dataset to generate a top-k matrix. similarity scores for the top-k matrix are calculated. one or more image of the dataset having fine-grained similarity with the search image are selected based a bundled similarity score for each image of the dataset. the bundled similarity score is a summation of the similarity scores of the image.



[[Category:G06V10/74]]
[[Category:G06F16/532]]
[[Category:G06F16/56]]
[[Category:G06V10/22]]
[[Category:G06V10/46]]
[[Category:G06V10/75]]