LawalMcCoy

Project Title

Improving subtitle visibility in low budget movies

Introduction

The main goal of our project was to develop a program in MATLAB that achieves visually pleasing viewing properties of a given subtitle and frame (picture) combination. Although the best solution is to have real people go through a movie, frame by frame, and decide what viewing properties for the subtitles makes them most visible, developing an automated solution may be the next best thing. In addition, many foreign films may not have the budget to have individual people review the movie as this is a tedious task.

The viewing properties of the subtitle that our program analyzed were the location of the subtitle, the shadowing around the text (if any), and the color of the text in the subtitle. For the color our program uses the CIELAB metric of deltaE as a proxy for the "correct" color. Color theory states that the larger the deltaE between two objects, the more discernible they are. In this situation we analyzed the average delta e difference between text and other pixels in the textbox.

One of our major goals was to answer the question: do larger deltaE differences appear more visually appealing to the viewer or do more subjective factors (such as shadowing and location) dominate? Furthermore, is there even really an advantage to larger deltaE differences (with all other factors held constant) in determining visually pleasing subtitles?

Method

How the program works

The program works through the use of a main wrapper function, place_captions. Before running the program though, preliminary work needs to be done. First the clips are converted into movie structs which are Matlab constructs that contain each frame in the movie as an image. Then, the subtitles must be transcribed into a text file in the same directory as place_captions. Now the program is ready to run. The program creates a black shadow around the text which research has shown increases discriminability between text and the image behind it. Next the location of the subtitle is so that it is in the lower left corner, and issues with this are discussed in the subsequent section. The program that we run finally selects the optimal shade of gray (equal amounts of RGB channel) that maximize the deltaE of the subtitles between the image the subtitle is overlaid on.

Results

test

alt text

Conclusions

test

References

test

Appendix I - Code and Data

test

Appendix II - Work Partition

test