search

Xiangtaokong / MiOIR

Towards Effective Multiple-in-One Image Restoration: A Sequential and Prompt Learning Strategy
62 stars 2 forks source link

question about the hazy image generation #2

Closed zyhrainbow closed 5 months ago

zyhrainbow commented 7 months ago

How to get the depth you've used. Thanks~

Xiangtaokong commented 7 months ago

We follow the commonly uesd Dehaze Dataset -- RESIDE. Specifically, the haze generation code is derived from "Learning Depth from Single Monocular Images Using Deep Convolutional Neural Fields" (github).

The citation number of this work in RESIDE is [38]. I will update this work to the reference of MiOIR to make it clearer.