SM Video Summarization

Had you ever downloaded a film from the internet and after days of downloading you found it was not the video you wanted .

Do you always have free space on your mobile to put a full film?

Can you watch the part from a film with put scrolling through the whole video?

Did you ever try to make a print screen for a shot and you got it as you want?

How long does it take to make a movie trailer and how mush you do pay for it?

Now you can forget this entire problem as SMSM-soft made the solution for it.

SM Video Summarization is a set of techniques for generating short summary for the content of long videos. Specifically, a video summary is a sequence of still or moving images representing the content of video. Which help customer to have a short meaningful summary of long videos this summary is coherent and consistent enough to cover the whole content of the video. It also saves user time and effort then making a manual summary or watching the whole video. Customers would be Web designers, Movie trailer makers, Mobile users, Sport program makers or even Video Librarians.

Options We afford

¨ General Summary:
This option gives a general overview to the most important shots in the video. This is made by calculating the importance of each shot and selects the most important shots to be displayed in new small video with less duration and size.

¨ Fast Summary:
This option generates photos from all over the video and the user will be able to save this photos with more the on extension.

¨ Time-based Summary:
This option allows user to enter duration of time and get a summary according to this duration. But this duration is limited by a range to have a meaningful summary.

¨ Background- based Summary:
This option generates photos of the different places in the video then when the user selects one of them the software generates a video of a shots happened in this place.

Video Summarization applications

· Web videos
Sometime web videos need to be summarized because users would like to have an idea of the content of a video listed on the web. When the user selects that summary, it can show the most useful frames. By this way, web users can view many summaries of videos in the same time that they would spend in watching one video, this video may be the undesirable one, so summaries help web users to decide which video will be watched or downloaded.

· VIPs
VIPs have no enough time to see long videos, although these videos are very important such as ones which record important meetings. If VIPs watch the summarized ones they will know what happened in these meetings without wasting time.

· Movie traliers
Video summarization can be used in media fields. If a user gets a summary for episodes or movies, this summary can be used for making movie intros and trailers which will save time and efforts for whom responsible of making them. It can also save money that is paid in this operation than the manual way

· Mobile
Wireless video phone can receive only three frames of video per second so video summarization will be helpful for these devices. Also it will save free spaces more than receiving the whole video.

· Video libraries
In libraries, readers take a look on the index of a book to know its field. If he is interested in it, he could decide to read it or not. In the same manner, summarized video will act as an index for its original video in libraries. Otherwise after he has watched the original video, he discovers that it is not what he wants. But this will happen after he is bored and spending a lot of time.

· Sports
Sport videos can be summarized depending on getting the interesting parts. This is useful in sport news.

· E-learning videos and personal uses
Also, video summarization may be used on personal computers, when the user wants to watch specific event on the video. For example assume there are a lecture has been recorded as a video and it has many events, the user want to watch specific one of them, so he will use video summarization system to extract this event which may be an experiment in laboratory, lecture explanation and so on. In addition, users can specify an object to make a summary about it like a person, a dog, a car…etc.


Challenges we met:

Conciseness is the first challenge; users would like to browse through the summarized video with a fast-forward that is smart enough to skip the dull moments. Also some devices such as wireless video phone can display videos with low rate. Users want to skip useless events and watch relevant ones. They have the patience to spend only 1% of the time it would take to stream the whole video so when users see the summary, they should feel it is like the original one without wasting time.

The second one is content coverage which gives the users general idea of the important contents of a video. After the summarized video is shown, users should feel that all events and actions in the original one are covered without discarding important ones.

And coherency. This summarization should preserve the meaning of the video. After summarization is finished, it should contain continuous events and understandable for the users.

Related work

Video structure analysis made by dividing the video into scenes; each scene into shots; each shot into images (frames); and then use image processing, and digital signal processing techniques. These techniques are used in summarization to select which frames will be accepted, and which will be discarded.

Object detection is another technique which is carried out to compare an object in a frame or shot with the rest of frames or shots to determine in which frames or shots this object was repeated.

Video shot detection is also a technique used in video summarization. A video shot is an image sequence captured continuously by a single camera. It is the basic building block of edited videos like movies, broadcast news, TV shows…etc.

There are many approaches for motion detection in a continuous video stream. All of them are based on comparison of the current video frame with one from the previous frames or with something that we'll call background. This technique called object tracking.

Tools used
· C# .net 2005
· SQL server 2005

Designed By : SMSM-soft