17377609. METHOD FOR RETRIEVING IMAGES FROM DATABASE simplified abstract (TAIWAN SEMICONDUCTOR MANUFACTURING COMPANY, LTD.)
Contents
METHOD FOR RETRIEVING IMAGES FROM DATABASE
Organization Name
TAIWAN SEMICONDUCTOR MANUFACTURING COMPANY, LTD.
Inventor(s)
Katherine Chiang of New Taipei City (TW)
METHOD FOR RETRIEVING IMAGES FROM DATABASE - A simplified explanation of the abstract
This abstract first appeared for US patent application 17377609 titled 'METHOD FOR RETRIEVING IMAGES FROM DATABASE
Simplified Explanation
Methods for retrieving images from a database are described in this patent application. The process involves obtaining a query image and extracting multiple patches from it. A set of weightings is obtained based on a sparsity-based dictionary, which includes non-zero weightings. The patches are then encoded using these weightings to create an encoding matrix. The database is searched using this matrix to retrieve images that correspond to the query image. Both the database and the dictionary can be dynamically updated to adapt to different query images encountered.
- Query image is obtained
- Multiple patches are extracted from the query image
- Set of weightings is obtained based on a sparsity-based dictionary
- Weightings include non-zero values
- Patches are encoded using the weightings to create an encoding matrix
- Database is searched using the encoding matrix to retrieve relevant images
- Database and dictionary can be dynamically updated to adapt to different query images encountered
Potential Applications
This technology can have various applications in the field of image retrieval and analysis. Some potential applications include:
- Content-based image retrieval systems
- Image recognition and classification algorithms
- Visual search engines
- Image recommendation systems
- Medical image analysis and diagnosis
Problems Solved
The methods described in this patent application address the following problems:
- Efficient retrieval of images from a database based on a query image
- Accurate matching of query images with relevant images in the database
- Adaptation to different query images by dynamically updating the database and dictionary
- Handling large image databases efficiently
Benefits
The use of these methods offers several benefits:
- Improved accuracy in retrieving relevant images
- Efficient and fast image retrieval process
- Adaptability to different query images
- Dynamic updating of the database and dictionary for better performance
- Potential applications in various fields including image recognition, medical imaging, and content-based image retrieval.
Original Abstract Submitted
Methods for retrieving images from a database are provided. A query image is obtained. A plurality of patches are extracted from the query image. A set of weightings is obtained according to a plurality of bases of a sparsity-based dictionary. The set of weightings includes a plurality of non-zero weightings. The patches are encoded with the set of weightings to obtain an encoding matrix. The database is searched based on the encoding matrix to retrieve the images corresponding to the query image. The database and the dictionary are dynamically updated to adapt to the query image encountered.
