Video quality and traffic QoS in learning-based subsampled and receiver-interpolated video sequences

Sources of real-time traffic are generally highly unpredictable With respec t to the instantaneous and average load which they create. Yet such sources will provide a significant portion of traffic in future networks, and will significantly affect the overall performance of and quality of service. Cl early high levels of compression are desirable as long as video quality rem ains satisfactory, and our research addresses this key issue with a novel l earning-based approach. We propose the use of neural networks (NN's) as pos t-processors for any existing video compression scheme, The approach is to interpolate video sequences and compensate for frames which may have been l ost or deliberately dropped. We show that deliberately dropping frames will significantly reduce the amount of offered traffic in the network, and hen ce the cell loss probability and network congestion, while the NN post-proc essor will preserve most of the desired video quality, Dropping frames at t he sender or in the network is also a fast way to react to network overload and reduce congestion. Our interpolation techniques at the receiver, inclu ding neural network-based algorithms, provide output frame rates which are identical to (or possibly higher than) the original video sequence's frame rate. The resulting video quality is essentially equivalent to the sequence without frame drops, despite the loss of a significant fraction of the fra mes. Experimental evaluation using real video sequences is provided for int erpolation with a connexionist NN using the backpropagation learning algori thm, the random NN (RNN) in a feed-forward cofiguration with its associated learning algorithm, and cubic spline interpolation, The experiments show t hat when more frames are being dropped or lost, the RNN performs generally better than the other techniques in terms of resulting video quality and ov erall performance, When the fraction of dropped frames is small, cubic spli nes offer better performance. Experimental data shows that this receiver-re constructed subsampling technique significantly reduces the cell loss rates in an asynchronous transfer mode switch for different buffer sizes and ser vice rates.