Smart Incontinence Status Alert Skill: Visual AI Automatic Detection of Wet Clothing and Abnormal Excretion

I Used an AI Skill to Solve the Biggest Headache in Nursing Homes

Last weekI volunteered at a nursing homeA caregiver complained to me: "The worst thing is when we don't notice an elderly person has wet themselves in timeBy the time we see the reddened skinit's already too lateThe elderly sufferand we feel guilty." To be honestthis is all too common in nursing homeshospital wardsand even at homeBedridden elderly peoplepatients confined to bedand babies—they can't speak or can't communicate clearlyBy the time caregivers notice something is wrongit's often too late.

But recentlyI discovered a super practical AI skill—the Smart Incontinence Alert SkillIt's essentially a visual AI-based recognition system that automatically determines from surveillance cameras or photos whether clothing is wet or whether there are abnormal excretionsand then immediately sends notifications to caregivers.

At firstI was a bit skeptical—can AI really tell if pants are wet? LaterI read the technical documentation and understood: it doesn't rely on guessworkIt works by analyzing image textureschanges in reflected lightand abnormal features in areas of body contactFor examplewet fabric shows different reflection effects under infrared or regular lensesand the system can precisely capture these subtle differences.

What Can It Actually Do?

In simple termsit gives caregivers a pair of "eyes that never blink." You can use it to analyze real-time surveillance videos or upload a patrol photo to check for abnormalitiesApplicable scenarios include:

• Nighttime care for bedridden elderly in nursing homes
• Regular rounds for bedridden patients in hospitals
• Wetness monitoring for babies during naps at home

Moreoverit doesn't just tell you "wet." It also generates a structured report recording the timelocationseverity of the abnormalityand even provides care suggestionsThis waycaregivers can handle it precisely without having to pull aside clothing to check every time.

How to Use It? Actually Quite Simple

This skill is open-source and can be found on GitHubIf you know a bit of Pythonyou can clone it directly and run itInstalling dependencies is just one line:

pip install requests>=2.28.0

Then enter the skill directory and execute the script:

python -m scripts.incontinence_alert_analysis --input /path/to/video.mp4 --user-type elderly --detection-mode real-time

If you don't want to bother with a local environmentyou can also use a cloud service directly—I tried running it with one click onand it was super convenient.

Pitfalls I Encountered

The first time I used itI foolishly uploaded a 200 MB videoand it took nearly 10 minutes to processLaterI realized the documentation stated: videos support mp4/avi/mov formatsmax 10 MBSame for images—jpg/png are finebut not too large.

Another timeI forgot to add the --user-type parameterand the system defaulted to "other," making the analysis less accurateLaterI changed it to elderly or infantand the accuracy improved significantly.

The most frustrating was checking historical reports—at first I was foolishly searching through local log filesfinding nothingLaterI learned I had to use the --list parameter to fetch data from the cloud API; local data is not stored.

Some Honest Thoughts

To be honestthis skill isn't perfectIt's sensitive to lighting and anglesIf the camera is blocked by a blanket or the light is too dimit might miss detectionsBut the flaws don't overshadow the merits; it's really helpful for improving care efficiencyEspecially for night-shift caregivers who manage multiple roomshaving this automatic alert system greatly reduces psychological stress.

I've also seen some nursing homes trying similar commercial solutionswhich cost tens of thousands of yuan per setBut this open-source skill is completely freeAs long as you have a camera and a computer (or cloud server) that can run Pythonyou can set it up yourselfFor small nursing homes or families with limited budgetsit's incredibly practical.

How to Get Started?

For those who want to give it a trygo directly to GitHub and search for smyx-open-claw-skillsthen find the incontinence-alert-analysis directoryIt contains a detailed SKILL.md document with illustrationsIf you don't want to read Englishyou can follow the steps I wrote above—they should be sufficient.

One last thing: No matter how powerful AI isit's only an aidTrue care still relies on human love and patienceBut this skill can at least help us reduce the regret of "not noticing in time," making the elderly and babies more comfortable.