티스토리 뷰
Improving Object Detection from Scratch via Gated Feature Reuse
Arc Lab. 2019. 8. 24. 01:28[업데이트 2019.08.24 01:16]
1. 논문
Improving Object Detection from Scratch via Gated Feature Reuse
2. 요약
- In this paper, we present a simple and parameter-efficient drop-in module for one-stage object detectors like SSD.
- We call our module GFR (Gated Feature Reuse), which exhibits two main advantages.
First, we introduce a novel gate-controlled prediction strategy enabled by Squeeze-and-Excitation [14] to adaptively enhance or attenuate supervision at different scales based on the input object size.
Second, we propose a feature-pyramids structure to squeeze rich spatial and semantic features into a single prediction layer, which strengthens feature representation and reduces the number of parameters to learn.







- Total
- Today
- Yesterday
- #TensorFlow
- Game Engine
- OST
- belief
- Jekyll and Hyde
- Worry
- Meow
- GOD
- project
- Memorize
- 도커
- Mask R-CNN
- ILoop Engine
- docker
- #ApacheSpark
- 2D Game
- Library
- English
- some time ago
- Ragdoll
- Physical Simulation
- aws #cloudfront
- Sea Bottom
- #REST API
- SSM
- #ApacheZeppelin
- ate
- sentence test
- Badge
- #ELK Stack
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |