Image Shows a Dolphin's Echolocated Impression of a Submerged Human • Hacker News Board

Re: Image Shows a Dolphin's Echolocated Impression of a Submerged Human

#21

Post by 100ideas » Wed, Dec 09, 2015, 8:40 AM UTC

100ideas wrote: post #20

porsupah wrote: post #15

I may just be being very slow today, but the methodology is rather unclear to me. eg "imprinted the signal onto a water membrane and then computer enhanced the resulting image" I'll admit, I'm.. not entirely persuaded by CymaScope's "biology" page: https://www.cymascope.com/cyma_research/biology.html "Cymatics--the trigger for life? Spiritual traditions from many cultures speak of sound as having been responsible for…

This page describes the methodology in a bit more detail: http://www.cymascope.com/cyma_research/oceanography.html According to that page, here is what they claim to have done: they placed an object (flower pot, plus-symbol, human body, etc) in front of a dolphin and recorded the echolocation sounds that reflected back from the object in the water as it was scanned by the dolphin. Then they reproduced the recording o…

Additionally, the US Navy has funded research into how dolphin sonar sensing and released a report in 1998 that demonstrated the generation of 2D images using "back projection acoustic tomography" of "synthetic dolphin signals". It's worth a read if you want to build your own DIY sonar system :)

I include it here because the work suggests dolphin echolocation pulses may contain sufficient information for dolphins to form something like a mental image of an object, although whether or not they actually do so is unknown. (They cite a 1995 paper by the lead author of the other paper I referenced, but not the potentially more-conclusive 1998 work).

"Tomographic Image Reconstruction of MCM Targets Using Synthetic Dolphin Signals", R. Altes, P. Moore, and D. Helweg. 1998. Accession Number: ADA337008, PDF Url: http://www.dtic.mil/cgi-bin/GetTRDoc?Location=U2&doc=GetTRDo...

"Extra information from elevation can be used to adaptively improve focusing and to construct a 3-D surface that represents the physical shape of an object as it would be perceived with vision. Such a surface is different from a representation of reflectivity as a function of position, and it allows direct comparison with visual representations (e.g., photographs) of objects. A visual analogue would be very useful to an animal that tries to perform sensor fusion by combining spatially registered feature maps from vision and echolocation. There is some speculation that dolphins may be capable of such a vision-like target representation (Pack & Herman, 1995), although the issue of the cognitive representation of targets formed by echolocating dolphins remains open to debate (Heiweg et al., 1996; Harley et al., 1995; Roitblat et al., 1995), and the lay concept of "seeing with echolocation" remains unsubstantiated. However, a vision-like target representation is well-suited to human observers and could have application to improvement of MCM performance."

The synthetic images are pretty neat. If you want to check them out but avoid the pdf, here is an imgur of the paper's figures: http://imgur.com/a/RLqF8

Re: Image Shows a Dolphin's Echolocated Impression of a Submerged Human

#22

Post by 100ideas » Wed, Dec 09, 2015, 9:12 AM UTC

100ideas wrote: post #20

porsupah wrote: post #15

I may just be being very slow today, but the methodology is rather unclear to me. eg "imprinted the signal onto a water membrane and then computer enhanced the resulting image" I'll admit, I'm.. not entirely persuaded by CymaScope's "biology" page: https://www.cymascope.com/cyma_research/biology.html "Cymatics--the trigger for life? Spiritual traditions from many cultures speak of sound as having been responsible for…

This page describes the methodology in a bit more detail: http://www.cymascope.com/cyma_research/oceanography.html According to that page, here is what they claim to have done: they placed an object (flower pot, plus-symbol, human body, etc) in front of a dolphin and recorded the echolocation sounds that reflected back from the object in the water as it was scanned by the dolphin. Then they reproduced the recording o…

I couldn't find any peer-reviewed publications by the researchers cited in the press releases by speakdolphin.com and cymascape.com - Jack Kassewitz and Jim McDonough et al.

However, I did find several scholarly articles that support the the hypothesis that dolphins seem able to recognize objects based on mental images they have formed via their echolocation sense, as opposed to recognition based on the object's particular texture, material, size, etc "signature". Here is one of them:

"Seeing through sound: dolphins (Tursiops truncatus) perceive the spatial structure of objects through echolocation." Herman LM1, Pack AA, Hoffmann-Kuhnt M.J. Comp Psychol. 1998 Sep; 112(3):292-305. http://www.ncbi.nlm.nih.gov/pubmed/9770316 PDF:[1].

From the paper's General Discussion section:

"The results also argue strongly that learning to associate particular characteristics of echoes returning from objects with the visual appearance of those objects was not the mechanism through which the dolphin in this study or the Pack and Herman study accomplished the cross-modal matching task (cf. Harley et al., 1996). Rather, the most compelling general explanation for the demonstrated cross-modal ability is that echolocation yields a direct percept of the object, in effect a representation of the object's shape that is analogous to or easily integrated with the percept developed through vision."

"In summary, the results reported here as well as those of Pack and Herman (1995) begin to construct a compelling case for a concept of echo imaging by the dolphin. The dolphin can apparently use returning echoes to construct a representation of an object that preserves the spatial structure of that object, to the extent that representation can be used to directly recognize that same object through the visual sense alone. The very short decision times used by the dolphin suggest that the object percept develops almost immediately and is of an integral whole, rather than limited to the examination and extraction of selected features."

[1]: https://www.researchgate.net/profile/Matthias_Hoffmann-Kuhnt...

Re: Image Shows a Dolphin's Echolocated Impression of a Submerged Human

#23

Post by bammmmm » Wed, Dec 09, 2015, 4:53 PM UTC

baxter001 wrote: post #4

Simply because it's possible to deconvolve a dolphins sonar clicks into a 2 dimensional representation approximating our visual senses doesn't mean they have any such similar experience of it, on top of that, the whole linguistic angle seems to be wild speculation.

jup, all they did was recording an impulse response of some click, which happened to be a dolphins.

Nothing about dolphin communication (I can't even find a reference in the press realease on the strange dolpin website linked in the article)

edit: I really hoped to lear, that dolphins speak in sonograms :)

Re: Image Shows a Dolphin's Echolocated Impression of a Submerged Human

#24

Post by dTal » Thu, Dec 10, 2015, 8:49 AM UTC

100ideas wrote: post #19

lobster_johnson wrote: post #17

This is spiritual new-age pseudoscience, and it's not worthy of HN. The web site is by a group called Sonic Age America [1], which turns out to be a guy named John Stuart Reid and his wife [2], whose occupation is "energetic healer". They believe that life started with sound, and have invented "cymatics, the study of visible sound" ("based on the principle that when sound encounters a membrane such as your skin or th…

I agree that much of their content has a "new-agey" feel, but that's not sufficient to discredit the particulars of the claims they are making about dolphin communication. It's conceivable their new-agey motivations for the work could nonetheless have led to a legitimate discovery, even if it is shrouded in quacky words like "cymatics".

"Cymatics" isn't quacky, though it may have been co-opted:

https://en.wikipedia.org/wiki/Cymatics