17971987. DETECTING FINE-GRAINED SIMILARITY IN IMAGES simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)
Contents
- 1 DETECTING FINE-GRAINED SIMILARITY IN IMAGES
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 DETECTING FINE-GRAINED SIMILARITY IN IMAGES - 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
DETECTING FINE-GRAINED SIMILARITY IN IMAGES
Organization Name
INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor(s)
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 17971987 titled 'DETECTING FINE-GRAINED SIMILARITY IN IMAGES
Simplified Explanation
Detecting fine-grained similarity in images involves identifying core areas of search images and comparing them with images in a dataset to find similar ones.
- Image salient map is generated from multiple layers of the search image to determine the core area.
- Feature descriptors are generated from the core area of the search image.
- Capsule vectors are generated from keypoints of the feature descriptors.
- Capsule vectors of the search image are compared with capsule vectors of images in the dataset to create a top-K matrix.
- Similarity scores are calculated for the top-K matrix to identify images with fine-grained similarity.
- Images with the highest bundled similarity scores are selected as similar to the search image.
Potential Applications
This technology can be applied in image search engines, content-based image retrieval systems, and image recognition software.
Problems Solved
This technology helps in accurately identifying images with fine-grained similarity, which can be challenging with traditional image matching techniques.
Benefits
The benefits of this technology include improved image search accuracy, faster retrieval of similar images, and enhanced image recognition capabilities.
Potential Commercial Applications
Potential commercial applications of this technology include image recognition software for security systems, e-commerce platforms for product recommendations, and visual search engines for online image searches.
Possible Prior Art
One possible prior art for this technology could be the use of feature descriptors and keypoints in image recognition and matching algorithms.
Unanswered Questions
How does this technology handle variations in lighting conditions when comparing images?
The technology does not explicitly mention how it accounts for variations in lighting conditions when comparing images. It would be important to understand if the system is robust enough to handle such variations in real-world scenarios.
Can this technology be applied to video content analysis as well?
The abstract focuses on image similarity detection, but it would be interesting to explore if this technology can be extended to analyze similarities in video content as well. This could have implications for video search engines and surveillance 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.