1. In the lecture of social multimedia computing, I have learnt that the multimedia can facilitate the computing of social networking. This is what I have not known before. I know that people upload their pictures to share with friends. Those pictures must contain some social information. But I have not thought to extract and use those information to construct a social network before. This is something that I have learnt in this lecture.
I also wish to learn further about how we can actively apply this skill in other field of usage such as social networking data mining and marketing.
2. I'd like to talk about the social multimedia search. Social Multimedia search is a way to first analyze the social multimedia, for example, first analyze a picture or a group of pictures. Then to extract and do processing of those multimedia, for example applying the SIFT algorithms to extract the representitiveness of the pictures. Then to group those with the high representitiveness in common and use it for certain ranking and categorization of the social multimedia.
I think one of the further implementation of this social multimedia search that does not cover in the lecture is to search for images of socially related characteristics. For example, we can search in SNS of friends pictures where we are shown in that pictures. Or we can search pictures of a particular friend. I think this usage will be very interesting. Actually, google has launched its imagery search of social circle that when we search on the web about some pictures, it can return with "Results from your social circle" to indicate that some pictures are from our album or our friends album. We can know more of this feature from this site: http://googleblog.blogspot.com/2010/01/search-is-getting-more-social.html#!/2010/01/search-is-getting-more-social.html
In the above link, a blog article is there to talk about this social media search enabled by google. There are also some examples in this article. I think this feature can facilitate people to interact more closely because very often, we do not know exactly what our friends are up to. And I think the social related search can return more accurate result.
While text search is already quite well developed, I also agree with you that a future area of implementation will be image related search. I like the idea of getting image search results from your social circle. Your online social circle are usually your real life friends or someone that has common interests with you. Returning images from them may be more relevant to your interests. This is just my opinion, not sure if you agree with me.
回复删除Yes, I agree with you, Andy.
删除I think the trend is that the search engine will try to utilize everything and every connections it can get to generate links and relations to facilicate the information mining. Social circle, I think, is a pool of personally centric information.
Still this application is under development and maybe in the near future, we can have easier way to find the information we want.
Best
Wang You
Actually for the search function of image, a huge database may be needed to built. That's is what you mean by "utilize everything and every connections" I think. Definitely this function is useful and will be widely accepted if it is realized.
删除I agree with your opinion that the further research on social networking searching for images of socially related characteristics. It is an interesting topic and very usefull for us to find some information about images. I am positive about the the opinion and I think it will be developed in the near future.
回复删除I have the same feelings about "I think one of the further implementation of this social multimedia search that does not cover in the lecture is to search for images of socially related characteristics."
回复删除You can read something and then never remember a single sentence. But if you see a great photo, you remember it. Make your business memorable with great images. No matter what your business is or your blog is on, images will only help you out. A person is smart, people are stupid. Cut them out of the crowd with pictures and then you can work on their higher thinking. Until then, attention is key. Grab them with a great image and then don’t let them go.
Well, I also know the great demand for image searching in online social platform. However, it is hard to implement. Why? Every image is required to contain enough information in order to be searchable. For example, add descriptions after each image uploaded. But generally most of us are too busy to add every image with detailed descriptions. It lead to uselessness of searching function. The best way to involve interaction is to highlight someone, just like the implementation in Renren.com.
回复删除