AI Landmark Recognition With Tensorflow Lite and CameraX on Android
Вставка
- Опубліковано 8 лип 2024
- In this video you'll learn how you can integrate a Tensorflow Lite model into your Android app to recognize landmarks with your camera by using the CameraX library.
💻 Let me be your mentor and become an industry-ready Android developer in 10 weeks:
pl-coding.com/drop-table-ment...
⭐ Courses with real-life practices
⭐ Save countless hours of time
⭐ 100% money back guarantee for 30 days
⭐ Become a professional Android developer now:
pl-coding.com/premium-courses...
Get my FREE PDF about 20 things you should never do in Jetpack Compose:
pl-coding.com/jetpack-compose...
Follow me on Instagram for insights into my live as an Android developer:
/ _philipplackner_
Download the landmarks AI model here:
tfhub.dev/google/lite-model/o...
Get the source code for this video on GitHub:
github.com/philipplackner/Lan...
At the end, I apparently forgot to show the screen mirroring, so you don't see what's happening. But, that shouldn't keep you back from trying it out on your own phone :)
Excellent strategy for developers to make the tutorial 😉😂
Philib you are great person, i am from somalia and you aspired me to learn native development
Real chad I bought your CI/CD course, even so I know a little bit but I love your work. Literally got me position partially so Thank Q for everything
Great ! , we need more AI videos, this will be great and will take the android development to the next level
As always, great content
Great person ! , we need more AI videos, this will be great and will take the android development to the next level
Great work as always ❤️🇨🇺
AI power 💪💪💪
Thank you for this amazing video .
Hey Philipp since you offered so nicely to do more videos on AI, could you kindly do one on MediaPipe like hand_landmarks or face contours. Many thanks for your great channel, its really a great resource for Android developers.
Amazing series. 🎉
I would be interested in the use of generative AI 😎 with VIT
Informative❤
I would like to see word processing, like adding an expense to a list and the model suggesting some categories for the user to add to the expense. Add an expense called Fuel and get some category suggestions, like car
Amazing content.
hello, first thanks for great videos. could you please explain how did you add the doted green frame on the camera live view ? and how is it possible to only focus on that area and take photos only from within the frame area ? thanks.
77th...Thanks Phillipp, Yes continue on the others like text, video and audio. Include also custom data training as well. Thanks in advance.
As an ML dev who is building an Android project, this is extremely helpful! However, it lacks the rendering of bounding boxes. Could you please make a video on that too? I am having issues scaling the boxes.
the best app it's!
hey, what animation did you use in video for "alt+enter" or something shortcut?
Love the tutorial, but I just don't agree with centerCrop part, you should've sacle down the image, cropping the image loses a lot of info about the image itself. I agree that scaling down also does loses info but don't forget that the model was trained on those scaled down images. And it is the technique used in the literature not cropping
great video as always! can you please make video on Text?
Hi Philipp, do you have an idea how to implement `setImageAnalysisBackgroundExecutor` so that no image preview is showing while analyze function is running?
Please. Tell more about this new assistent bot that will apear in in Android Studio
My biggest question in compose-ios, where do I see the compatibility of klotin modules transpiled to Swift?
Hello can we use this tutorial to build a native module and later use in our react native application?
Hello! Something wrong with youre github link - there is just default jetpac compose app from android SDK......
can you provide a tutorial for mediapipe and android
Can you do a video on how to run llms on Android please. It will be so usefull
Hi Is It Possible To Display The Output Of A Model That Gives An Image Output?.
hey bro Good day can you make a video on how to protect all .XML file and Res files in Gradle Version 8.1 ResGuard doesn't work for this version.
hi @Philipp! I am big fan of you your content is just amazing. Can you please tell me which .tflite model can be used for text recognition, I already used Mlkit but result is not satisfying, sometimes it detects text and sometimes not, so I want to use .tflite models for text recognition so that they can detect custom text. Please guide.
Hello Philippe, the github repo seems to be empty
how did you learn such stuff?
Thank you, the project compiled and classifies the images, but for some reason the square frame with a dotted line along the contour is not displayed. What could be wrong?
I didn't show that part in the video as it wasn't really related to the video topic
@@PhilippLackner How to make a photo frame
Detecting paper and crop it. please
I am confused because I train ml model custom. How to integrate it with android ? I already save model to .tflite but it seems that I need some metadata to deploy it on android
Bro I learning android programming since 6 month but I can't understand well or it's very hard, how can I keep self study?
0:46 how do i see my phone in my laptop as well. What software is this?
Webcam
💓
Android Studio showing error notification never changes anywhere.
hey I'm getting a error message for val imageProcessor = ImageProcessor.Builder().build() as Unresolved reference: Builder does anyone know as solutions to it.
i was having the same issue, check the import import androidx.camera.core.ImageProcessor.Builder
if same then remove it and then go on this line val imageProcessor = ImageProcessor.Builder().build() imageProcessor will be red in suggestion there i will be options choose tensorflow one if not solved even then go for other option mine is sovled
I tried using a custom tflite model instead of downloading an existing model, the application didn't run as expected as it didn't detect any objects, I rechecked if the tflite model was working fine, and it did but I couldn't connect it to this application, can I get some insights on it, like should I have a label.txt file, can anyone who know about this help me a bit
Hello, did you manage to get it work with your custom model? I’m also about to do something similar
@@chriskipalamoto2811 I continued using pre-trained models and noted that if you switch to a custom-trained model, you might need to adjust your code, that might help you access data from a text file listing all the image names.
ua-cam.com/video/zs43IrWTzB0/v-deo.htmlsi=9Vg-R5abhU2uhqwj
You might find this video helpful, it includes codes for recognizing objects with a custom TensorFlow Lite model. With small adjustments you'll get the code to work like the one shown in this video
@@shahdiyamaheesha2509 thank you! I will give it a try
There is no code on github link. Main activity is empty
APP folder missing in source code
I literally copy and pasted and the app keeps crashing when started😢
i have the same problem
🪄🪄