Abstract: Errors in dialing or otherwise entering network access codes such as telephone numbers are identified by comparing each entered code with a set of stored codes. Each entered code is also stored. For each stored code, statistics are compiled that reflect the likelihood that the code is invalid. Such statistics include frequency of use over a given time period, the length of the connection to the network after each use, and the length of time after ending a short connection that a different but similar or more probably correct code is entered. Similarity between codes is preferably measured by determining in how many positions the codes agree. If an entered code has a statistically low likelihood of being correct, then the system replaces it with a code that is sufficiently similar, or immediately follows it sufficiently often. The system checks, for example, for possible likely transposition errors.

Abstract: Reflectors such as microbubbles in a contrast agent introduced into a blood vessel of a patient are destroyed (or otherwise acoustically altered) using focused ultrasound at a modulation point according to an input modulation sequence. This creates “gaps” in the flowing contrast agent that are sensed at a downstream sensing point. The pattern of gaps is then matched in time with the input modulation sequence to determine a transit time for the gaps, which is also the flow velocity of the blood. The input modulation sequence creates at least two gaps, but may otherwise have any of several different forms, which include, among others, square-wave, maximal sequence, random binary patterns. Edge-detection and correlation techniques are used to match the input and sensed gap patterns. By triggering the input sequence off of a heart rate monitor, a flow velocity profile may also be calculated and displayed from one heart beat to the next.