딥러닝 훈련소

  • 홈
  • 태그
  • 방명록

re-sampling 2

[논문리뷰] Instance-Aware Repeat Factor Sampling for Long-Tailed Object Detection(IR

https://arxiv.org/abs/2305.08069 Instance-Aware Repeat Factor Sampling for Long-Tailed Object DetectionWe propose an embarrassingly simple method -- instance-aware repeat factor sampling (IRFS) to address the problem of imbalanced data in long-tailed object detection. Imbalanced datasets in real-world object detection often suffer from a large disparity inarxiv.org 1.Introduction안녕하세요 저번 포스팅에서 W..

논문 리뷰 2024.05.26

[논문 리뷰] Dynamic Curriculum Learning for Imbalanced Data Classification

1.Introduction저번 포스팅에서와 같이 long tail 데이터셋에서의 예측 성능을 높이기 위한 기법을 제시된 논문입니다. 차이점은 이 논문에서는 ImageNet과 같은 데이터셋과는 달리 CelebA 데이터셋을 중심으로 설명하고 있습니다.(RAP와 CIFAR-100 데이터셋으로 실험한 결과도 제공합니다.) CelebA 데이터셋은 여러 유명인들의 얼굴 이미지인데, 총 202,599개의 얼굴 이미지로 구성되며 10,177명 가량의 사람들(명당 20장 정도)이 데이터셋을 이루고 있습니다. 주목할 점은 각 사진마다 40개의 특성 정보들(Smiling, Male, Young, Big nose,...)이 있고, 1또는 0으로 라벨링 되어있다는 것입니다. 따라서 이 데이터셋은 각 사람에 대한 샘플 수는 균형..

논문 리뷰 2024.04.10
이전
1
다음
더보기
프로필사진

딥러닝 훈련소

대학원생의 딥러닝 공부방입니다~

  • 분류 전체보기 (27)
    • On-Device AI (1)
    • 딥러닝 CV&LLM (7)
    • 딥러닝 기초이론 (10)
    • 논문 리뷰 (7)

Tag

인공지능, repeat factor, learning rate, re-sampling, vanishing gradient, 손실함수, 딥러닝, irfs, long-tail, Computer Vision, 컴퓨터비전, Classifier, 시그모이드, cnn, 경사하강법, Object Detection, gradient descent, 선형회귀, Ai, CV,

최근글과 인기글

  • 최근글
  • 인기글

최근댓글

공지사항

페이스북 트위터 플러그인

  • Facebook
  • Twitter

Archives

Calendar

«   2025/08   »
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
31

방문자수Total

  • Today :
  • Yesterday :

Copyright © Kakao Corp. All rights reserved.

티스토리툴바