Xiangyu-Fu / TUM_Computer_Vision

the challenge of the computer vision course

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

TUM_Computer_Vision

It is the repository of the final challenge project for Computer Vision (SS22) at Technical University Munich.

Lecturer: Prof. Dr.-Ing. Klaus Diepold

Team Member:

  • Qiwen Xu,
  • Xiangyu Fu,
  • Xiaodong Lei,
  • Yunyu Chen,
  • Yuntian Yan.

Introduction

"Tout Into the Picture" allows users to observe the scene through a virtual camera by building a 3D scene to achieve the effect of a tour in the picture[1].

We develop a Matlab app. In this app, users can observe the scene presented by the selected picture from different perspectives, initially achieving the same effect as the paper.

Here is a exmaple of our project,

2022-07-18-23-56-09 mp4_20220718_235809


GUI Environment

cover

💻 : work both for Windows 11 and macOS Monterey

🚀 MATLAB Version: at least MATLAB R2022b and Toolbox Image Processing Toolbox must be installed

⌨️ GUI develop kit: APP designer


How to use our GUI

Step 1

First, open the mian.m and click the run botton, image

Then the GUI will be opened like this, 1658177597339

Click the Select Dataset here button and select the folder which you store your images. image

Here we use the example dataset as the input folder and we select the simple_room.jpg as our Image Tour example.Once you selceted the image, the next step will occur. After that you can begin your tour of image!:smile: image

Step 2

In the second step, you will be taken to the foreground selection page. Please note that if you do not have a foreground in your image, please skip this step directly using the Nest Step button. image

If there exist some foregrounds in your selected iamge, please click the Select Foreground button and mark the area in the foreground of the image. Our GUI allows you to select multiple foregrounds. image

Depending on the size of the image, the processing of the image takes different amounts of time. After the process, you will see the image showing like this. Then you can click the Next Setp button to continue your tour. image

Step 3

In this Step, Please sclect the Vanishing point and the rear wall to help our algorithm to build the 3d coordinates, please use Select Vanishing point and Select rear wall seperately. image Our GUI will display vanishing points and vanishing lines to help you check if your selection is correct. If it is not correct, you are free to reselect the vanishing point or the rear wall; if you think there is nothing wrong, then please use the Next Step button to enter the picture tour. image

Step 4 : Tour into the image

In this page, you can begin your room tour. Please use the slides in the right-hand side to move you position. Have fun! If you want to try a new image, please use the Select a new iamge button. image tour

About

the challenge of the computer vision course


Languages

Language:MATLAB 100.0%