Hi Maninda, I agree with the concern.
However, as far as the Hemika's project is concerned, lets accept that inefficiency to keep it within 3 weeks scope. Having said that, I am very much interested how video stream like data set can be received and processed though an event stream (within Siddhi). May be the best we can do is to process them at an input adapter and only send high level events. If you understand this domain, can we chat about this a bit. --Srinath On Thu, Nov 13, 2014 at 6:17 PM, Maninda Edirisooriya <[email protected]> wrote: > In general video streams are compressed in time domain. If we send one > frame as an event this will not be achieved and the required bandwidth will > be higher. > Instead if the video stream can be chunked and decompress inside the > server / engine, bandwidth could have been saved. > > BTW we can use the same input adaptors to get videos to BAM and do some > batch processing if map-reduce implementations can be found for computer > vision algorithms like http://hipi.cs.virginia.edu/ > > > *Maninda Edirisooriya* > Senior Software Engineer > > *WSO2, Inc.*lean.enterprise.middleware. > > *Blog* : http://maninda.blogspot.com/ > *E-mail* : [email protected] > *Skype* : @manindae > *Twitter* : @maninda > > On Thu, Nov 13, 2014 at 8:32 AM, Srinath Perera <[email protected]> wrote: > >> Adding dev@ >> >> Hi Suho, Mohan, >> >> Open CV (computer vision library) let you train and build a model that >> generates an XML file that can be used from a java program to classify >> video streams (e.g. detect an human, vehicle). >> >> Idea of the project is to write an CEP extension that takes the XML file >> and detect vision detection. We can send each frame in video as an event. >> However, it is a byte array. How can we send this? If he used siddhi, can >> he send it by defining type as object? >> >> Long term we should define an byte[] >> >> --Srinath >> >> On Wed, Nov 12, 2014 at 3:35 PM, Hemika Kodikara <[email protected]> wrote: >> >>> Hi all, >>> >>> This is regarding the Fast Track Training Project that I selected to >>> complete [1]. >>> >>> The task is to create a CEP extension which will receive a stream of >>> video and do several image processing tasks to identify an object. >>> >>> Need to know more insight on how to send input a video stream. >>> >>> Even though it is a video, we can use it frames. Which will be a set of >>> images. >>> >>> [1] - https://redmine.wso2.com/issues/2646 >>> >>> Hemika Kodikara >>> Software Engineer >>> WSO2 Inc. >>> lean . enterprise . middleware >>> http://wso2.com >>> >>> Mobile : +94777688882 >>> >> >> >> >> -- >> ============================ >> Blog: http://srinathsview.blogspot.com twitter:@srinath_perera >> Site: http://people.apache.org/~hemapani/ >> Photos: http://www.flickr.com/photos/hemapani/ >> Phone: 0772360902 >> >> _______________________________________________ >> Dev mailing list >> [email protected] >> http://wso2.org/cgi-bin/mailman/listinfo/dev >> >> > -- ============================ Blog: http://srinathsview.blogspot.com twitter:@srinath_perera Site: http://people.apache.org/~hemapani/ Photos: http://www.flickr.com/photos/hemapani/ Phone: 0772360902
_______________________________________________ Dev mailing list [email protected] http://wso2.org/cgi-bin/mailman/listinfo/dev
