Method and apparatus for analyzing sensor data
Mobile devices, such as smartphones, are severely battery-limited. The capabilities of mobile device processors have increased at exponential rates, but battery technologies have improved at much slower rates. As a consequence, it is ever more important that mobile devices be operated in battery-preserving manners. An aspect of the present technology concerns methods and arrangements enabling on-going recognition-processing of imagery and audio by different detectors, without the high battery drain that has characterized the prior art. A variety of other features and arrangements are also detailed.
1. A method comprising the acts:
(a) receiving plural sets of sensor data from a sensor;
(b) evaluating whether each of the received sets of sensor data likely has a signal that can be extracted;
(c) reporting results of said evaluations to a marshal module;
(d) for a first of said received sets of sensor data, which is evaluated as likely having a signal that can be extracted, the marshal module instructing an extraction module to attempt extraction with a first level of intensity;
(e) for a second of said received sets of sensor data, which is evaluated as likely having a signal that can be extracted, the marshal module not instructing the extraction module to attempt extraction; and
(f) for a third of said received sets of sensor data, which is evaluated as likely having a signal that can be extracted, the marshal module instructing said extraction module to attempt extraction with a second level of intensity different than said first level of intensity.
2. The method of claim 1 in which act (e) is based, at least in part, on the extraction module successfully extracting a signal from a previous received set sensor data.
3. The method of claim 1 in which act (e) is based, at least in part, on a computational budget that would be exceeded by such attempted extraction.
4. The method of claim 1 in which said plural sets of sensor data comprise audio or image data.
5. The method of claim 1 in which the plural sets of sensor data comprise image data, and in which the evaluating act comprises evaluating whether said image data likely has a digital watermark signal or a barcode signal that can be extracted.
6. The method of claim 1 in which said evaluating comprises generating a probability datum indicating an estimate of whether an extraction module will achieve a positive result if applied to one of said sets of sensor data.
7. The method of claim 1 in which the plural sets of sensor data comprise image data, and in which act (e) comprises the marshal module instructing a digital watermark decoder, a barcode decoder, a face recognizer, an optical character recognizer, a logo recognizer, or an image recognition engine, to attempt extraction from said image data.
8. The method of claim 1 in which the plural sets of sensor data comprise imagery, and in which act (d) comprises the marshal module instructing a digital watermark decoder to attempt extraction, wherein the watermark decoder is configured to examine different image blocks within said imagery, or image blocks with different scales or perspective distortions within said imagery.
9. The method of claim 1 in which said evaluating is performed using a neural network.
10. The method of claim 1 that includes limiting processing resources employed by said extraction module so that a previously-established processing budget is not exceeded.
11. The method of claim 1 that includes the marshal module choosing between said received sets of sensor data that are evaluated as likely having a signal that can be extracted, to select said first set of sensor data for attempted signal extraction by the extraction module, said choosing being based in part on history data about previous processing of sensor data by the extraction module.
12. The method of claim 11 in which the history data includes data indicating an elapsed interval, either of time or intervening sets of received sensor data, since a previous processing of one of said previously-received sets of sensor data by an extraction module.
13. The method of claim 1 that includes the marshal module determining said second level of intensity based, at least in part, on a history of previous signal extractions from previously-received sensor data.
14. The method of claim 1 that includes the marshal module determining said second level of intensity based, at least in part, on a quality of said third received set of sensor data.
15. The method of claim 1 that includes the marshal module determining said second level of intensity based, at least in part, on an orientation of a device of which the sensor is a component.
16. The method of claim 1 that includes the marshal module determining said second level of intensity based, at least in part, on a motion of a device of which the sensor is a component.
17. The method of claim 1 in which said evaluating yields numeric data for each received set of data, said numeric data indicating a relative likelihood that the received set of sensor data has a signal that can be extracted.
18. The method of claim 17 that includes the marshal module determining said second level of intensity based on both (a) said numeric data, and one or more of: (b1) a history of previous signal extractions from previously-received sensor data, (b2) a quality of said third received set of sensor data, (b3) an orientation of a device of which the sensor is a component, or (b4) a motion of said device of which the sensor is a component.
19. A non-transitory computer readable medium containing software instructions for configuring a processor-equipped device to perform acts on plural received sets of sensor data including:
(a) evaluating whether each of the received sets of sensor data likely has a signal that can be extracted;
(b) reporting results of said evaluations to a marshal module;
(c) for a first of said received sets of sensor data, which is evaluated to likely have a signal that can be extracted, the marshal module instructing an extraction module to attempt extraction with a first level of intensity;
(e) for a second of said received sets of sensor data, which is evaluated to likely have a signal that can be extracted, the marshal module not instructing the extraction module to attempt extraction; and
(f) for a third of said received sets of sensor data, which is evaluated as likely having a signal that can be extracted, the marshal module instructing said extraction module to attempt extraction with a second level of intensity different than said first level of intensity.
20. A system for operating one or more signal extraction modules, comprising:
first means for evaluating whether each of plural received sets of sensor data likely has a signal that can be extracted; and
second means for determining, from among first, second and third received sets of sensor data evaluated by the first means to likely have a signal that can be extracted, that (a) extraction of a signal should be attempted from the first set of sensor data by a signal extraction module with a first level of intensity, and that (b) extraction of a signal should not be attempted from the second set of sensor data by a signal extraction module, and that (c) extraction of a signal should be attempted from the third set of sensor data by a signal extraction module with a second level of intensity different than said first level of intensity.