Reproduce the paper ๐ฅ"An Edge-Less Approach to Horizon Line Detection"๐ฅ
The code is used to detect the skyline(the boundary between the sky area and the no-sky area) of the images.First, generate the 16*16 patches as train dataset from the mountain images; Second, train a classifier based on SVM; Third, slide on the test image to get the probability of patches which belong to skyline.
- SVM
- data generate
- support train and test
- python3.6
- numpy==1.18.1
- opencv==3.4.2
- scikit-learn==0.22.1
@INPROCEEDINGS{7424466,
author={T. {Ahmad} and G. {Bebis} and M. {Nicolescu} and A. {Nefian} and T. {Fong}},
booktitle={2015 IEEE 14th International Conference on Machine Learning and Applications (ICMLA)},
title={An Edge-Less Approach to Horizon Line Detection},
year={2015},
volume={},
number={},
pages={1095-1102},}