Originally Posted by JD-EE
This brings to mind the many eyes approach several groups are using to find interesting bits in huge volumes of data. If the bottom profiles were released to the public for processing and scrutiny both the state of the art and the chance of discovery of "bits" in large amounts of data improve.
This is a great idea, except that I'm concerned that it would not produce useful results. In sidescan sonar records, a lot of things may stand out as targets, and may even have angular shapes that look like man made objects, but when investigated, they turn out to be rocks. If there were any large parts (wings, fuselage section, etc.) they would likely have been spotted by the eyes that already looked at the data. Even if there might be smaller parts in the data, the number of "false alarms" generated by inexperienced eyes would likely swamp any ability to investigate or even discard them.
I expect that the searchers are looking for patterns of targets that might be different from the surrounding rock fields. They likely got their local experience by investigating some of the early clusters and discovering they were rocks. We know that the ROV was used to investigate some targets. It seems likely that they learned what not to waste time on. That sort of knowledge is difficult to pass on. (Of course the flip side is that something ignored could have been debris.)
This search is not like the Fossett search, where color images can be scanned for white metal objects in a desert setting. The sonar images are composed of just varying reflection strength. While it's not as bad as looking at long range radar data that is just little blips and clutter, in some ways it is more like that than looking at a photo.
Perhaps someone will comment who has more experience than me in reading high clutter sidescan images.