TODO

我们的pixel和cm对应关系；我们有多少normal patch，>60才行吧（参考MVTec AD中的最小数据集）？

[database + 双模态无监督配准block] + AD

reconstruction-based AD: masked convolution block

1 Introduction

1.1 What

Untitled

1.2 Why

1.2.1 Is it object detection, can be done by YOLO [11], SSD [12])?

it is infeasible to collect all possible FOD examples in the dataset representation because of the open-ended nature of FOD. Limitations of the dataset could cause FOD detection systems driven by those supervised algorithms to miss certain FOD, which can become dangerous to airport operations.

Dataset

images

  1. FOD-A: A Dataset for Foreign Object Debris in Airports, 21

    https://github.com/FOD-UNOmaha/FOD-data

    for detection and classification, not for location!

    用AD来做，差什么？

  2. Foreign Object Debris Detection for Airport Pavement Images based on Self-supervised Localization and Vision Transformer, 22

    https://github.com/fod-unomaha/fodanomalydata

    for location.

    The localization method requires a separated training/validation set which contains FOD-free images of runways, and a testing set which contains images that include FOD randomly distributed around the image

References

  1. Foreign Object Debris Detection for Airport Pavement Images based on Self-supervised Localization and Vision Transformer, CSCI 22

    1. FOD based on Self-supervised Localization and Vision Transformer, CICS 22

  2. SLiDE, eccv22