18045448. ARTIFICIAL INTELLIGENCE APPARATUS AND METHOD FOR DETECTING UNSEEN CLASS ITEMS THEREOF simplified abstract (LG ELECTRONICS INC.)
ARTIFICIAL INTELLIGENCE APPARATUS AND METHOD FOR DETECTING UNSEEN CLASS ITEMS THEREOF
Organization Name
Inventor(s)
ARTIFICIAL INTELLIGENCE APPARATUS AND METHOD FOR DETECTING UNSEEN CLASS ITEMS THEREOF - A simplified explanation of the abstract
This abstract first appeared for US patent application 18045448 titled 'ARTIFICIAL INTELLIGENCE APPARATUS AND METHOD FOR DETECTING UNSEEN CLASS ITEMS THEREOF
Simplified Explanation
The present disclosure describes an artificial intelligence apparatus and method for classifying and detecting items of an unseen class that have high visual similarity to items of a known class.
- The apparatus can classify and detect items that have not been previously learned or known.
- The method involves encoding and decoding item images using a codebook of encoded data for each item class.
- The unknown item is detected by classifying the class of the item image based on the decoded data.
Potential applications of this technology:
- Object recognition: The apparatus can be used to classify and detect objects in images, even if they belong to an unseen class.
- Fraud detection: The method can be applied to identify fraudulent patterns or activities that have not been previously encountered.
- Medical diagnosis: The apparatus can assist in identifying and classifying medical images, even for conditions that have not been previously studied.
Problems solved by this technology:
- Overcoming limitations of traditional classification methods that can only detect known classes.
- Addressing the challenge of classifying items with high visual similarity to known classes.
- Enabling the detection of previously unseen or unknown items.
Benefits of this technology:
- Improved accuracy in classifying and detecting items, even for unseen classes.
- Enhanced ability to adapt to new and evolving data.
- Potential for discovering new patterns or classes that were previously unknown.
Original Abstract Submitted
The present disclosure relates to an artificial intelligence apparatus capable of classifying and detecting an item of an unseen class having high visual similarity to an item of a seen class that has been learned and known in advance, and a method for detecting an unseen class items thereof, and when the item image is input, generate encoded data by encoding the item image, generate decoded data by decoding the encoded data using a codebook of a plurality of codebooks in which encoded data for each item class is stored, corresponding to a pre-learned item class, and detect the unknown item by classifying the class of the item image based on the generated decoded data.
