this post was submitted on 09 Jan 2024
186 points (94.7% liked)

cats

19299 readers
1933 users here now

typical internet cats. videos, pics, memes welcome!

rule 1) be kind

lemmy.world rules:

other cat communities midwest.social cats

founded 1 year ago
MODERATORS
 

Hello everyone! I am looking for cat owners keen to help out with a research project, I am studying cats to see whether we can estimate their age just by their voices. πŸ—£οΈπŸˆ For this, I need data. Lots of it.

It's quick and simple - please help me by recording your furry friend's meows according to the specific instructions in the image below.* πŸ“πŸŽ€ Any contribution is invaluable and will help gain insight into the development of age-related vocalisation patterns in cats! 🧠🐾

The more meows the better, so please feel free to send me multiple recordings! πŸŽ§πŸ“ˆ

*n.b. By sending this data, you are consenting to its use for research purposes as part of a deep learning tool for age estimation of the domestic cat. This study aims to develop a non-invasive method to determine a cat's age through vocalisations. This approach promises significant benefits in veterinary care, can aid rescue centres in creating accurate adoption profiles, and has potential implications for understanding the age demographics of feral cats.

** EDIT:

Adding some additional info following a user comment

This is part of my dissertation project for my MSc in AI, I am working on building a pipeline for age prediction based on vocalisations. While the primary goal is to achieve a robust deep learning model for age estimation in cats, this project could serve as a pipeline for other species in future. I ultimately want to work in the field of bioacoustics and use the sounds of nature for work in conservation efforts, this is the first step towards that.

Depending on how the project goes, I would like to build an app so that this can be used in the field by veterinarians and rescue centres and the such. Of course this would include a whole lot of extra GDPR and real-time processing efforts. If time allows it I might build one for demonstration purposes and release it in the future.

I will release the implementation details in full upon completion. I am working through comparative development of certain architectures which I think will be useful for the wider field.

I do not have a conflict of interest. I recommend Dolby On to record the meows as it allows for Waveform Audio File format (.wav) recordings that I require for processing and feature extraction, and easy exports.

Here is a copy of the email that each participant receives, detailing data subject rights and right to withdraw at any time:

Any other questions, just let me know! I really appreciate any input and participation, thanks :)

all 34 comments
sorted by: hot top controversial new old
[–] [email protected] 23 points 10 months ago (1 children)

I have a 2 year old cat who sounds like she's been smoking for 40 years, may not be the best representative. ;)

[–] [email protected] 20 points 10 months ago

Hey, that’s good data! There are always outliers, right? My 12yo meows the most and definitely sounds a bit like an old lady; I’m trying to remember if she always did. One of my 4yos has a super girly high-pitched voice. I struggle to imagine what she’ll sound like in 10 years.

[–] [email protected] 21 points 10 months ago (1 children)

What organization are you researching with? Why is the submission via Gmail?

How are you handling participant data (mostly email addresses, it seems)? Can participants opt out and revoke access to the data after submission?

Do you have a conflict of interest? Do you or any of your colleagues have an affiliation with Dolby or other companies involved in the research?

Did this proposal pass IRB? I'm guessing it's exempt, so probably yes, but do you have the approval number?

What do you plan on doing with the model? Are these data for training the model or for testing it?

I know those questions sound a lil aggro, and to be clear, I don't think there are necessarily right answers. Maybe you're an undergraduate or hobbiest, like... I don't think IRB is super important for a cute cat study. But I do think this kind of info should be included in recruitment calls as a standard!

Cheers, seems cute and fun

[–] [email protected] 19 points 10 months ago* (last edited 10 months ago)

Hello, rightful questions - apologies for not including this!

I'm an MSc student in AI and this is for my dissertation. I did consider a Google form or similar, however via email seemed easiest considering the Dolby app has a direct share option (I am not affiliated no, but I do think it is an excellent app). In terms of approval, I received ethical approval from my university to go ahead with my research proposal.

Every participant will receive an email with the further points you mention. I'll attach a copy here.

And thanks for the heads up, I will include this info when I post elsewhere.

[–] [email protected] 17 points 10 months ago

I'm in. When do you send me my cat?

[–] [email protected] 16 points 10 months ago (1 children)

Mine doesn't even meow, she only makes the cat activation sound, but all the time. Especially when running and jumping.

[–] [email protected] 8 points 10 months ago

Cat activation sound, love it.

[–] [email protected] 11 points 10 months ago (1 children)

Shouldn't you ask for the cat's age as well as the recordings? Or did I read over that?

It might be worthwhile to gather some other information about the cat, like whether they're indoor or outdoor cats, former strays or not,...

[–] [email protected] 10 points 10 months ago (1 children)

Second to last paragraph in the first image - "don't forget to tell me your cats' age along with the recording(s)"

[–] [email protected] 2 points 10 months ago

Thanks! I knew I must have missed it, but couldn't find it.

[–] [email protected] 9 points 10 months ago

CatGPT folks

[–] [email protected] 7 points 10 months ago (1 children)
[–] [email protected] 6 points 10 months ago (1 children)

Why is it showing removed on lemmy.world? Does .ml or .world censor bitch or did you type remove yourself Aight, test: bitch

[–] [email protected] 8 points 10 months ago (1 children)

LOL .ml censors the word bitch.

[–] [email protected] 5 points 10 months ago

Yeah, it seems there's an automated profanity filter which I was not aware of until recently. I wish you could toggle this on and off as a user, but hey, I'm a guest on this platform so who am I to removed about this, am I right?

[–] [email protected] 6 points 10 months ago (1 children)

How accurate does the age need to be?

[–] [email protected] 9 points 10 months ago (2 children)

I don't require the exact age, but the age would need to be certain on the year as I will be grouping the cats into age brackets [kitten, young adult, adult, senior].

[–] [email protected] 5 points 10 months ago (1 children)

You may wish to reconsider this. At the moment you're excluding everyone who adopted an adult cat of poorly documented past; we're probably not that many, but we tend to want to be helpful and often own seniors. Vet assessment can usually narrow age range down to abt 3 years.

[–] [email protected] 5 points 10 months ago

I'll take it on a case by case basis, if a cat is certainly over say 10 years old I can classify it as a senior but if it could be 3 or 7 it is more tricky.

[–] [email protected] 4 points 10 months ago

Thanks for the response and good luck with the study!

[–] [email protected] 6 points 10 months ago (1 children)

My 17+ year old does seem to have a slightly raspier voice than she used to.

[–] [email protected] 3 points 10 months ago (1 children)

Please record her for me if you can! I don't have many seniors yet.

[–] [email protected] 2 points 10 months ago (1 children)
[–] [email protected] 1 points 10 months ago

Got it, thank you so much!

[–] [email protected] 6 points 10 months ago (1 children)

Sounds like a cool project! I have some questions just out of curiosity.

  • What type of model are you considering? Would it potentially fit in mobile devices? Would love to open an app or smth after meeting a random cat in the street to estimate its age lol.
  • How many recordings are you planning to work with? I don't have experience with audio but with NLP and many other domains NNs require huge amounts of data to work correctly. What are you thinking to mitigate this issue?
  • Do you plan to open-source the model or notify us again when its ready somehow?

Good luck with the project! My cats very rarely meow but if I get to record them I'll send it to ya.

[–] [email protected] 5 points 10 months ago (1 children)

Thank you, I am very excited to be working on this.

I would like to keep the full development details under wraps for now, but I will of course release everything in full once the project is finished! I would love it if people could use my research for their work. The idea is to build an app but this depends on the accuracy that I will achieve and time limit.

I shall report back to you in a few months!

[–] [email protected] 4 points 10 months ago

Very understandable, hope it works out!

[–] [email protected] 4 points 10 months ago (1 children)

Good luck! Chubby cat does not meow, and the kitten sort of...whines instead of meowing.

[–] [email protected] 6 points 10 months ago (1 children)

Thank you! Whines could be useful, if you manage to capture any feel free to send it over :)

[–] [email protected] 2 points 10 months ago (1 children)

I've realized a question!

He won't make any sound if I don't start him off. It's it okay if audio of me chatting with him is in there?

[–] [email protected] 2 points 10 months ago

That's ok! I will isolate the meows, so that's fine :) my cats are the same

[–] [email protected] 3 points 10 months ago (1 children)

I already have an app that can record lossless audio, is that fine, or is there something special on the Dolby app?

[–] [email protected] 4 points 10 months ago

That's fine! I have included the dolby app as standard phone recorders generally don't seem to export in true lossless, and compress it. The Dolby app allows export in Waveform Audio File format (.wav) which I process the data in, but if you have something similar that is fine too.