Title:
MODEL TRAINING APPARATUS, MODEL TRAINING METHOD, AND COMPUTER READABLE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2022/254597
Kind Code:
A1
Abstract:
An object of the present disclosure is to provide a model training apparatus, a model training method, and a non-transitory computer readable medium capable of providing time efficient method to determine the optimal Feature Pyramid Network (FPN) count parameter value for an Artificial Convolutional Neural Network (ACNN) model. In one aspect, a model training apparatus (100) includes an estimation unit (101) configured to estimate a value of the number of FPN blocks included in an ACNN model by using a training dataset, and an ACNN model training unit (102) configured to train the ACNN model by using the estimated value.
Inventors:
VAGHANI DARSHIT (JP)
Application Number:
PCT/JP2021/020948
Publication Date:
December 08, 2022
Filing Date:
June 02, 2021
Export Citation:
Assignee:
NEC CORP (JP)
International Classes:
G06N3/04
Domestic Patent References:
| WO2017168922A1 | 2017-10-05 |
Other References:
GOLNAZ GHIASI; TSUNG-YI LIN; RUOMING PANG; QUOC V. LE: "NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection", ARXIV.ORG, 16 April 2019 (2019-04-16), pages 1 - 10, XP081169738
HIRAKAWA TSUBASA, YAMASHITA TAKAYOSHI, FUJIYOSHI HIRONOBU: "Architecture Search for Distant Object Detection", 1 May 2020 (2020-05-01), pages 1 - 6, XP093009366, Retrieved from the Internet [retrieved on 20221219]
HIRAKAWA TSUBASA, YAMASHITA TAKAYOSHI, FUJIYOSHI HIRONOBU: "Architecture Search for Distant Object Detection", 1 May 2020 (2020-05-01), pages 1 - 6, XP093009366, Retrieved from the Internet
Attorney, Agent or Firm:
IEIRI Takeshi (JP)
Download PDF:
Previous Patent: SEMICONDUCTOR DEVICE
Next Patent: ROTARY SCREEN PRINTING DEVICE AND ROTARY SCREEN PRINTING METHOD
Next Patent: ROTARY SCREEN PRINTING DEVICE AND ROTARY SCREEN PRINTING METHOD