Invention Grant
- Patent Title: Joint training of neural networks using multi-scale hard example mining
-
Application No.: US18462305Application Date: 2023-09-06
-
Publication No.: US12154309B2Publication Date: 2024-11-26
- Inventor: Anbang Yao , Yun Ren , Hao Zhao , Tao Kong , Yurong Chen
- Applicant: Intel Corporation
- Applicant Address: US CA Santa Clara
- Assignee: Intel Corporation
- Current Assignee: Intel Corporation
- Current Assignee Address: US CA Santa Clara
- Agency: Hanley, Flight & Zimmerman, LLC
- Main IPC: G06V10/00
- IPC: G06V10/00 ; G06F18/243 ; G06N3/04 ; G06N3/08 ; G06V10/44 ; G06V10/82 ; G06V20/10 ; G06V20/70 ; G06V30/19 ; G06V30/24

Abstract:
An example apparatus for mining multi-scale hard examples includes a convolutional neural network to receive a mini-batch of sample candidates and generate basic feature maps. The apparatus also includes a feature extractor and combiner to generate concatenated feature maps based on the basic feature maps and extract the concatenated feature maps for each of a plurality of received candidate boxes. The apparatus further includes a sample scorer and miner to score the candidate samples with multi-task loss scores and select candidate samples with multi-task loss scores exceeding a threshold score.
Public/Granted literature
- US20240013506A1 JOINT TRAINING OF NEURAL NETWORKS USING MULTI-SCALE HARD EXAMPLE MINING Public/Granted day:2024-01-11
Information query