Novel Self-Portrait Enhancement via Multi-Photo Fusing

APSIPA, Dec 2016.

• Sifeng Xia          sfxia18@163.com
• Shuai Yang        williamyang@pku.edu.cn
• Jiaying Liu         liujiaying@pku.edu.cn
• Zongming Guo  guozongming@pku.edu.cn

 

Introduction

In this paper, we present a novel multi-photo-based framework to solve a self-portrait enhancement problem we call “1+2 problem”, in which a self-portrait photo is enhanced with the help of 2 multiple photos that share the same scene and similar shooting time. The key idea is to exploit the extra information of these 2 photos to overcome the limited field of view and poor illumination of the target self-portrait. Given 1+2 photos, our method first performs an illumination adjustment to accomplish a uniform lighting style. Then these photos are aligned using the homography transformation matrices which are estimated by RANSAC over the matched SIFT feature points. Furthermore, a Markov Random Field based approach is proposed to globally fuse the aligned photos nicely.

Experimental Results

The proposed self-portrait enhancement method is implemented on Visual Studio 2013 platform. In the experiment, we set λ = 1 and α = 2. To the best of our knowledge, there does not exist such a system that focuses on self-portrait enhancement and can achieve both unified illumination and expanded FOV. The closest one is the Photoshop PhotoMerge tool [1]. Thus in the experiment, we compare the performance of the proposed method with Photoshop. The self-portrait photos and the corresponding supporting photos for testing are taken by iPhone 6 Plus.

Experiment 1:

We show the results of diffentent sets of raw photos in Fig.1~4. In each set, the first row shows the raw self-portrait and supporting photos, the second row shows photos of llight uinification and Photoshop and proposed results are shown in the third row.

Experiment 2:

We show the performance of the proposed method on expanding a large field of view of the test set shown in Fig.3. Fig.4 shows the improved result.

References

[1] Adobe: Photoshop PhotoMerge, http://helpx.adobe.com/en/photoshop/using/create-panoramic-images-photomerge.html

Citation

@INPROCEEDINGS{xia2016se,
author={S. Xia and S. Yang and J. Liu and Z. Guo},
booktitle={Asia-Pacific Signal and Information Processing Association Annual Summit and Conference},
title={Novel self-portrait enhancement via multi-photo fusing},
year={2016},
pages={1-4},
month={Dec}}

Back to Projects Page

---