Optimal Grouping-of-Pictures in IoT video streams

We study a dynamic video encoder that detects scene changes and tunes the synthesis of Groups-of-Pictures accordingly. Such dynamic encoding can be applied to infrastructures with restricted resources, like IoT facilities where multimedia streams are of use. In such facilities the scarcity of resources (energy, bandwidth, etc.) is a dominant solution design factor. In the domain of video capturing/transmission content-driven approaches should be adopted to improve efficiency while maintaining quality at acceptable levels. We propose a time-optimized decision making model that yields different sizes of groups-of-pictures (frames) to meet the previously discussed objectives i.e., transmit video sequences in acceptable quality with rational use of the wireless resources. Our quantitative findings show that the propose scheme performs quite efficiently while dispatching video sequences with different characteristics.