Automating My Life with Python | Using Computer Vision to Detect How Often I Drink Coffee

[@edson_winner]

Thank you. You just gave a glimpse of how object detection works. Kinda thinking of doing something similar in mining industry to detect different kind of rocks and sizes.

[@lkgpuanimho0349]

Someone in Indonesia did something like this, using OpenCv to detect sex, age and emotions. Then the door behinds him opened and OpenCv detected a 500-year old angry female ghost. True story.

[@redx001-xxx]

👀

[@KakashiHatake-jl9gx]

i got interest in tech cause of you...and i'm loving the process learing new things !!

[@TiffInTech]

that made my day! Thank you so much :)

[@KakashiHatake-jl9gx]

😄

[@theophilus494]

one thing i like about you is your sincerity on how you get things done.. awesome..

[@kumardigvijaymishra5945]

Perhaps adding a piece of code that could train the machine to detect different kinds of cup would do the trick. But the cup needs to be detected with open mouth. Cool project Tiff.

[@Flat0Line1]

For this kind of stuff - it's like jump from medieval age to the landing on the moon, sadly. Here was just used default classifier for face detection, and actually doesn't really solve the task of "coffee drink detection" and any other custom detection thing. But as first time with openCV it's ok..

[@kumardigvijaymishra5945]

@@Flat0Line1 instead of face detection, the code needs to detect the cup, any cup. 'Coffee drink detection' will require a pressure sensor on the coffee mug, NFC device that could send signal to the code, ... that would be landing on moon 🌝

[@ozono27]

@@kumardigvijaymishra5945 or detecting the cup AND that the cup has been tilted while the mouth disappears, kinda? without any doubt... the sensor on the mug is much better... for sure.

[@shaunweinberg2463]

Full stack dev moving to Data science (again) and doing honorus in Physics maths. I drank up to 20 cups of coffee at a time. I had to change my diet due to the caffeine causing stomach issues. Keep the coffee and sugar down...

[@Flat0Line1]

Interesting idea for a project!)) You have a problem with testing on video because your white_pixels are not normalized by the size of the face in the camera view (essentially, it doesn’t check the distance between the face and the camera). So, if you are a little further from the camera, white_pixels will be around 8000-9000 or even 5000, if you are a little closer, it can be more than 80,000.

[@ozono27]

might be a good idea to create the threshold not on the amount of pixels in the "region of interest" but the ratio between the pixels of the ROI and the pixels of the face, right? that way, distance would top being an issue? Of course,... looking sideways will create a problem always.

[@espnmk]

Pretty Cool Tiffany thanks for Sharing, definitely will be playing around with this. Your such a wealth of knowledge.

[@Nitin-wo5xp]

Keep going 🎉

[@TiffInTech]

Thank you!!

[@pandipatipavan3804]

​@@TiffInTechPlease do provide the code files and the links in the description as soon as possible for free 🙏 thank you so much ❤️ All the Best 👍

[@UberGeek]

Cool idea, but it is clear the program is detecting the "white" from the lettering on the coffee mug instead of your teeth, as you have explained. Especially since you are NOT opening your mouth before you lift the mug to... your mouth.😜 How does the code distinguish the difference when you are talking? A better solution might be to have the code identify the "Mug raising" to your mouth. ☕

[@HeyItsJonSchwartz]

code review (and maybe video review) suggestion: If you are trying to explain and/or remember what an acronym means and it's in code you are unfamiliar with, write it out. mouth_roi should be mouth_region_of_interest. It may be tricky to put up on screen, but we're not paying by the character in python, so it's better to make it dead stupid to remember what it meant later. Side effect is you don't need to waste video time explaining that to us. Also, github link?

[@TiffInTech]

Thanks for sharing!!

[@HeyItsJonSchwartz]

@@TiffInTech I just reread what I wrote and I hope you take that comment the way I intended which is to say trying to be helpful but happy to have you tell me to kick rocks if you disagree. I really liked the video and I hope you share more like this.

[@Flat0Line1]

Yep, it's very good approach in general - at work especially)

[@moisesespiritosanto2195]

Hi, I'm from São Paulo! Yeah excellent! ABRAÇO FORTE DE URSO!

[@TiffInTech]

Yay!! Hope you have an amazing weekend!

[@moisesespiritosanto2195]

​@@TiffInTechHey, you too!!!

[@oooooohmylisa]

你很好看，好喜欢你哈哈哈，视频不错

[@aresaurelian]

Well done, Tiff! We love when you explore the unknown (to most of us) of AI. This is so exciting and interesting. What a fantastical world we are in. And together is more fun.

[@Sumandass_1]

Lots of love from India

[@TiffInTech]

Wishing you all the best!

[@PointEndClick]

This video is awesome.

[@TiffInTech]

Thank you!💗

[@TheMetalMag]

excellent thank you! you give different view on creating project

[@Jibril_Abdulkadir]

This is why I code to do funny projects 😂😂

[@TiffInTech]

Hahah totally

[@rudyvaldez]

Bad ass!

[@TiffInTech]

Cheers!!

[@othmanaljbory3649]

I'm love you thank you so much

[@TiffInTech]

Cheers!

[@DavidTheWiseOne]

@@TiffInTech Lucky lucky chair 🪑 😍😍

[@G_a_n_d_u]

oye oye balle balle

[@AvinashSingh-vj3rk]

❤

[@TiffInTech]

💗💗💗

[@WilsonMar1]

You have a github repo of code for what you're showing?

[@thgorobets]

Oh, well, I need a cup of Coffee right now!

[@chukwuemekailozor9105]

Did the code confirm that it was coffee in the cup? What if you were drinking whiskey?

[@DrJ_Endoking93]

Lovryou babe❤

[@chiluone]

Nice Background Music :D :D

[@Ilovecookingtheinternetbums]

Hi Tiffany, please what exactly do you? What's a day life in Tiffany's like?🤣. I think ppl would be very interested. Thanks.

[@TiffInTech]

Forsure!!

[@davidmckee533]

Baddd tiff tooo much coffee!!! the next level should be a warning you have entered "kidney stone level". something to that effect. Thanks Tiff you are great

[@TiffInTech]

Haha! Promise it’s not actually that much 🫣😆 thank you!!

[@LinuxForLife]

I think you are wrong when trying to detect your "open mouth"... 😉 I think it will be more accurate to focus on the cup (of coffee) detection.

[@MyExRebirth]

If you need AI vision to control and remind you of your cofee drinking habit, you might wanna use a catheter during work as well

[@kriscatz2298]

Wondering if it is possible to make a mobile app that does this? 🤔

[@Civernet00]

I confirmed she's pretty good

[@B_knows_A_R_D-xh5lo]

try hot chocolate! 🎉❤😊

[@TiffInTech]

Yessss! I love chocolate so I should

[@steveheyden9219]

How many cups at a time？

[@TiffInTech]

always too many cups!!!

[@aryansahare1145]

I usually count myself

[@MTMixTV]

From India 🇮🇳🪷

[@TiffInTech]

Wishing you a great day!!

[@MTMixTV]

@@TiffInTech Here It's Night 😂

[@JustinMacri007]

I wish tiff was my friend

[@ronnie-codes]

I envy how healthy you look.

[@JimDooley]

haha. say "th" 10 times fast.

[@TiffInTech]

😆😆

[@romulomanrique6765]

eres demasiado hermosa

[@kennethshupe8946]

I want to see this combined with Pavlok, that way instead of a notification, you would get shocked. Perhaps if someone is trying to quit vaping.

[@TiffInTech]

Haha ohhhh!! I mean maybe that would help people

[@KonstantinKovalenko-d3b]

How often I drink coffee🤦‍♂can't you count?

[@mrbane2000]

This video made me little annoyed, because its not working well. And i'm not even Python i ditched it in 2016 when i started learning it because i though it will be popular. But i found my self fixing indentions when copying code annoying. Although there is Bython now. So i made my own version how i think this code would work better....

[@mrbane2000]

Just put face detection also and if face is gone and circle is near face condition to improve it. Quite easy, every cup has a circle... import cv2 import time as time import numpy as np cap=cv2.VideoCapture(0) inicialCircles=0; sip = 0 inicialTime=int(time.time()) timeSinceSip=0 while True: _,frame=cap.read() actualTime=int(time.time()) timeSinceSip=actualTime-inicialTime image = frame.copy() gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) circles = cv2.HoughCircles(gray, cv2.HOUGH_GRADIENT, 1.2, 100) if circles is not None: circles = np.round(circles[0, :]).astype("int") for (x, y, r) in circles: if r

[@tony9146]

😂 going full CCP on yourself This is great though. You could supplement this with a scale that measures the weight of the cup so that you can measure rate of consumption, and you could even place a QR code or barcode on your cup.

[@shaun.august]

Python is great for automating things and managing data, but you don't need Python or computer vision just to figure out how many times you drink coffee. We can drink coffee up to 5 times at most, no need to count it. That doesn’t make sense. computer vision should be used for things like tracking how many times you toss and turn at night while sleeping. that makes a complete sense.

[@zainabbas6167]

tiff the best in tech 🫶 Thanks for sharing these kind of videos 😊

[@TiffInTech]

Thank you!! That means a lot!

[@zainabbas6167]

​@@TiffInTech you are welcome 🫶