Read on Twitter
I am looking for how they created a data set for the "analprints" but in the meantime, here is an article about it that features actual butthole photos; you are welcome. https://www.vice.com/en_us/article/jge377/stanford-smart-toilet-uses-butthole-for-identification
They are calling the combo of fingerprint and analprint, "two factor user identification" and I will never see that term again and not think of butts.
The figures feature a drawing of a child standing to pee and an actual turd photo.
This paper is a gift.
“11 participants were recruited for system testing over a period of 5 weeks.” How? How did they recruit them? Did they put up flyers? Are they enthusiastic friends and family? Did they bribe graduate students with pizza? If they did would that influence results?
“Two board certified surgeons (sub specialty of coloproctology) independently annotated all the stool images with a corresponding BSFS to assign them as the gold standard.”
It looks like the data set is only 11 butts? "ani from 11 research participants were analysed and the anus morphologies were compared (Stanford IRB approval...). Among 11 participants, 2 video clips of the anus per participant were acquired from 7 participants..."
Yes. I think so. "As an input, individual frames of the anus from participant 1 were used for identification purposes. We performed 10,201 comparisons among video-extracted frames for each participant...301 frames from the video were extracted and fed to the CNN.
Side note: Gender did not affect how the poo looked, but survey results showed that males were more accepting of having their buttholes photographed.
They are nothing if not thorough. Honestly this is one of the better poo studies I have read and I like to think I have read more than many.
This gem: "classifying one’s own stool is sometimes unpleasant and requires extra effort for the participant, hampering reliable self-reporting over an extended period of time."
In the future: "For quantitative stool analysis, the anal scanner...will be used to monitor the entire defecation processes....The expected result is a precise description of hydrodynamics of human defecation"
Reading SI now: They did use Google! "Our original dataset contains 2,902 stool images spanning all 7 BSFS types..These stool images were acquired through search engines such as Google, Bing and Yahoo (keywords included ‘stool in toilet’ and ‘faeces’).
They retrained Google's open source neural net on poo and anus images online + those collected from their participants.
"As real stool images from research participants accumulate, the performance of our CNN will be considerably enhanced."
They did use flyers! "participants were recruited through flyer advertisements and e-mail lists. Once participants showed interest, a short introduction of the research study was provided, along with possible risks and benefits..."
Data is stored in DropBox. Not sure how I feel about that.
This is a thing I did not know->"opioid-induced constipation is also a frequent clinical problem faced by clinicians"
They used Go-Pro cameras, as one does.
Oh! They also made a smartphone app! Of course they did. "The current version of the application allows users to upload a stool image acquired by the users and the transferred tensors classify the stool accordingly."
Here is the corresponding figure.
What if your butt changes? They are on it: "The aspect of the anus may change over time. Regarding the morphological stability of the anus over time, a few studies have already dealt with this concern, by describing automatic template update strategies for biometrics."
That concludes my live-tweeting of this fascinating poo and anus paper. If there are any editors out there who would like an actual article covering this research, featuring probably the most interesting interviews ever, feel free to reach out. This is in fact my wheelhouse.
