BROWSE
Detail
Registered METHOD OFTRAININGIMAGEDEEP LEARNING MODELANDDEVICETHEREOF
- Inventors
- Cho Yong Ju, Jeongil Seo, 모흐센 알리, 리한 하피스, 무하마드 파이살, 아만 이샤드
- Application No.
- 16879885 (2020.05.21)
- Publication No.
- 20200372350 (2020.11.26)
- Registration No.
- 11720790 (2023.08.08)
- Country
- UNITED STATES
- Project Code
- 19ZR1100, Development of Fundamental Technology for Hyper-Realistic Media Space, Jeongil Seo
- Abstract
- Disclosed herein is an image deep learning model training method. The method includes sampling a twin negative comprising a first negative sample and a second negative sample by selecting the first negative sample with a highest similarity out of an anchor sample and a positive sample constituting a matching pair in each class and by selecting the second negative sample with a highest similarity to the first negative sample, and training the samples to minimize a loss of a loss function in each class by utilizing the anchor sample, the positive sample, the first and second negative samples for each class. The first negative sample is selected in a different class from a class comprising the matching pair, and the second negative sample is selected in a different class from classes comprising the matching pair and the first negative sample.
- KSP Keywords
- Learning model, deep learning(DL), loss function, negative samples, training method
- Family
| Status | Patent | Country | KIPRIS |
|---|---|---|---|
| Registered | METHOD OF TRAINING IMAGE DEEP LEARNING MODEL AND DEVICE THEREOF | KOREA | KIPRIS |