Dynamic identify algorithm for H.223 protocol identifier based on neural network

For the validity of the H.223 protocol identifiers, the paper deeply analysed the H.223 multiplexing protocol and the main influence factors of extraction, and proposed an extraction algorithm based on BP neural network, which could adjust the judge threshold based on the current communication circumstance in real time. Simulations show that the proposed algorithm can effectively enhance the accuracy rate compared to the ones based on the static threshold extraction, so that improves the video quality of 3G mobile equipment under the circumstances with high bit error rate.