This proposal is about the design of device which would be used to identify the selected / viewed TV channel, duration of viewing and to generate the reports on the audience-ship of a TV channel, and eventually help in understanding the TV viewership pattern.  TV viewership patterns are critical in determining the ratings and popularity of the channel hence driving the commercial impact. The product is designed to be connected to the AV out of the television set, which by default gives the composite video signal of the currently selected / watched TV channel. This composite signal would be periodically captured by the device and subjected to a classification algorithm and output parameters such as currently viewed TV channel, time of viewership, etc are recorded. The classification process would try to find the logo, which appears at corners of the screen. Once the logo is identified and captured, it will be compared with the logos database of the channels. Viewership metering record is marked and data is dispatched to the viewership-pattern cloud for analysis reports. This device would be equipped with SIM enabledGSM module, which will be used as medium of communication. Once a channel is identified; Name & ID of the viewed channel duration of view, date, time and location of the device will be transmitted to the viewership-pattern cloud. Application hosted on the cloud will store the data and help generate viewership reports against time and TV channel.

It is presumed that such a product would be able to detect the logos with in a good precision and be able to give data regarding the viewership pattern of the TV channels. Establishment of viewership pattern with degree of confidence can solve the problem of rating issues being faced right now in Pakistan.

The most advanced methods being used in the world correlates the audio samples at the user site with the one in the data base of the server. The server already has a delayed audio copy of all the stations being broadcasted. The received samples from the user site would be then matched with the server data set and a decision is made. Such a method may need a reasonable amount of bandwidth and would have a reasonable operational cost, however if the matching process is done over the user device, then operation cost could be minimized. Proposed schemes ensure that all shortcomings are managed efficiently at the user premises equipment and ensure crowd-sourced viewership patterns.