In a perfect follow-up to yesterday's article on natural language processing that used HAL as an example, today's news involves an update to Intel's well-known OpenCV computer vision library. The new software provides audio-visual speech recognition (AVSR); the ability to read lips. By combining lip-reading with traditional audio language recognition, much lower error rates are achieved. A new LinuxDevices article describes the software. A recent Slashdot article also covers the AVSR software and includes the usual lively discussion. The OpenCV library is Free Software released under the GPL-compatible Intel Open Source License.