单目法向估计算法介绍

任务

输入一张单目RGB图像，单目法向估计算法将分析场景三维结构、输出图像对应的稠密法向图

模型描述

本模型基于Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scans算法。

技术细节请见：

[Project Page] | [Paper]

如何使用

代码示例

import cv2
from modelscope.outputs import OutputKeys
from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks

model_id = 'Damo_XR_Lab/cv_omnidata_image-normal-estimation_normal'

input_location = 'data/test/images/image_normal_estimation.jpg'
estimator = pipeline(Tasks.image_normal_estimation, model=model_id)
result = estimator(input_location)
normals_vis = result[OutputKeys.NORMALS_COLOR]
cv2.imwrite('result.jpg', normals_vis[:, :, ::-1])

输入输出示例

输入：

输出：

适用范围

默认输入图片的分辨率为384x384, 非此尺寸的图片会resize到这个分辨率

Bibtex

@inproceedings{eftekhar2021omnidata,
  title={Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets From 3D Scans},
  author={Eftekhar, Ainaz and Sax, Alexander and Malik, Jitendra and Zamir, Amir},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={10786--10796},
  year={2021}
}