Realtime Object Detection Android App | Kotlin | Camera2 api | tflite

Hi in this video you will learn to build an android app which can detect object and draw rectangle around it in real time.
We will use Camera2 api to read frames from camera and will use tflite tensor flow lite to detect for objects in the frame and finally canvas to draw the rectangle on the frame.
Hope you will learn something new today.
After this video you will know:
1 - How to use Camera2 api
2 - How to preview the processed frame
3 - How to deploy tensor flow lite model
4 - How to setup TextureView
This is a computer vision android app which is built by using kotlin.
model link : www.tensorflow...
labels.txt : gist.github.co...
code: github.com/Paw...
contact me :
facebook : / proogramminghub
twitter : / programming_hut
github : github.com/Paw...
discord : / discord
linkedin : / programminghut
youtube : / programminghutofficial
android studio,kotlin,computer vision,tensor flow lite,object detection,realtime,camera2 api,texture view,surface texture,bitmap,canvas,draw rect,mobile net,tflite

[@gustavomartins1247]

I have been trying to work with flutter packages for quite a while now with no success, I tried modifying my code on kotlin side following your video and it worked, thanks a lot!

[@MuhammadUmar-ul7op]

i just copy and past the code and it's work for me thanks

[@MrAzzu41]

Thanks sir successfully completed this course and add it to protfolio

[@yashwanthmamaduru7086]

Can you send your project?

[@4ia10_muhammadriyan6]

can you give an example for the yolov8 model?

[@ssrpic]

Thank a lot brother. The project absolutely works!

[@yashkhopkar3126]

Bro can you make a seperate video on object detection with java full video step by step

[@mohomedashfak5514]

thanks for the video. it really helped me ❤️❤️

[@codingtamilan]

Bro... please do it slowly or make tutorial for this mostly machine learning or Deep learning people's don't know about any Java or kotlin language

[@rimurudota]

i trained a model but it has no outputs locations classess scores numberofdetection. only val outputs = model.process(inputFeature0) val outputFeature0 = outputs.outputFeature0AsTensorBuffer please help

[@AkshayS-us4jq]

THANK YOU SOO MUCH BRO ,THIS HELPED ME A LOT, LOVE YEW LOTS.......

[@glennimmanuel9338]

is it work for you? cause in my app it cannot detected the object

[@TheMatrixWarriors99]

Het brother that a really nice turtorial but i want to know which model you have used for object detection.

[@sainathingle]

can you make tutorial for object recognization or detection for image using cloud api like google vision or any other api

[@frankieyeung7064]

great video, but how to fix the textureview aspect ratio deform problem?

[@alanood9500]

I have a question about YAMNet TensorFlow lite model (Android app). I want to use it with an audio clip as input, Not a live recording. Can you help in that? Thank you for your help

[@vijaygarg8034]

"Nice video! Could you please guide me on how to save images to the gallery and perform predictions on real-time camera input and still images?"

[@amirfaizm]

Hi, I use Android API 34 for the sdk platform,these 3, SDK build tools, Android Emulator and sdk platform tools for the sdk tools. and yet, when i try to run the code (the code has no error), it says "please select android SDK". Do i need a specific sdk tools to run the code?

[@mikebledig7208]

Is it possible for you to modify the code of this app to use voice output to speak the object the cam sees? Could you show us how we would go about doing that using tts (text to speech)

[@marianstoikov8218]

Can you do it on Java am new to your channel love your videos really good understanding

[@exor6748]

this is giving me error when I am using my own custom tflite model...I have also changed the labels but its still throwing error..please help

[@bingebox4070]

My model output shape is [1,6,8400] , so how can I get coordinates class labels and confidence score from output= Outputs.outputAsCategoryList?? Please Help me on this Brother , I looked online and could'nt find anything.

[@kukilp213]

I am getting blackscreen on emulator (Pixel Xl Android 12) and Redmi Note 9 pro. It was working fine till camera feed display. Please check. Error Log: E/CameraManagerGlobal: Camera 61 is not available. Ignore physical camera status change.

[@puranjitsingh1782]

Great video, also can you let me know how can we generate the tot number of detections as an output from the model (generating results for predictions of different classes)

[@user-uz6rx9fg2k]

Hey did you find a way?

[@lykagracetabiolo7418]

Great i think this is the one video that you suggest me to want for deploying yolov5 in android.❤ do you have a java version of this code? so that it will be compatible with the yolov5 model I have deploy earlier from your other video

[@programming_hut]

Process will be similar you can try working out using chatgpt, maybe you can ask it to convert the code to java

[@dargotime6609]

nice, can we detect objects and record them in video?

[@davizpoh]

super thanks...

[@wmusicstudio945]

thank you sir

[@mohammed._.sameer7753]

I am getting blackscreen after introducting of model section what should ido

[@muhammadhussnain61]

sir i have a problem related to emulator.. pixels problem sir.. emulator screen is blurry

[@AlsibarJAhmad]

can you please make a video tutorial about this but add a voice speak that says what is the camera detecting like for example apple then the apps will say "apple" please

[@yokijatiperkasa861]

thanks sir

[@fansdrone5752]

The video is great. Many thanks. 👍👍 I follow your video to deploy my Gundam Recognition tfile to android app. 👍👍 But i got a problem, when i rotate my phone to landscape, the detection position will not correct. I had already written a rotation function but i got problem as i can't get textureView and camera bitmap in aligned rotation (only at portrait is aligned). Can you teach me how to solve? Many thanks.

[@sushantkandwal]

Can you please send the code

[@ToddMay-z9c]

Rather than using a physical cameras device... Does anyone have experience using this with a FFMPEG YUV feed? Or some other grabbing of an existing video stream and feeding it into this code?

[@studymeow525]

Can we add a currency dataset to tflite model? To make object and currency detector both together?

[@rajveervaidya3656]

Bhai aapne bnaya ka yeh please bataye

[@karanjitrandhawa1515]

Unresolved reference: Builder Hi sir I am facing this issue. Can you please help me out

[@maximood-tired]

hello, great video. what is your font in Android Studio?

[@programming_hut]

It’s chalkboard

[@maximood-tired]

@@programming_hut Apple Chalkboard or something else? Do you know how to get it on windows or linux?

[@programming_hut]

I’ve never done that but I think this might help stackoverflow.com/questions/31681167/using-custom-font-for-android-studio-editor You can probably download chalkboard if possible or try any other alternative

[@maximood-tired]

​@@programming_hut alright thanks👍

[@sushantkandwal]

@@programming_hut can you provide full code I am getting penty of error

[@LAJianoran]

how can i delay the result without affecting the camera?

[@isnaayu_m]

great, how if i use .ptl ? can I ?

[@NameName-jk4py]

Unable to instantiate activity ComponentINfo Getting this error at 10:24

[@richlejesta3359]

val inputFeature0 = TensorBuffer.createFixedSize(intArrayOf(1, 224, 224, 3), DataType.FLOAT32) inputFeature0.loadBuffer(byteBuffer) sir what should i put in byteBuffer? or change]

[@devpk1325]

same here

[@devpk1325]

val model = Yolov5s.newInstance(context) // Creates inputs for reference. val inputFeature0 = TensorBuffer.createFixedSize(intArrayOf(1, 416, 416, 3), DataType.FLOAT32) inputFeature0.loadBuffer(byteBuffer) // Runs model inference and gets result. val outputs = model.process(inputFeature0) val outputFeature0 = outputs.outputFeature0AsTensorBuffer // Releases model resources if no longer used. model.close()

[@ahmadsarji]

It works for me but i am asking how i can add voice output feature to output the detected object??!!

[@shreyatalveda1792]

Did you do that?

[@sergosokolov4678]

Why don't have ready app for download?

[@MightyStrongGuy]

i love you 😍

[@AshishMaley]

meko labels.txt ni milri bhai

[@ngsingwong3083]

hi sir, what macbook system are you using?

[@programming_hut]

Macbook air m1

[@steel-r_ua]

Please, do this: *Ctrl + A, Ctrl + L*

[@UPS_]

i can't find the code

[@IqraZahid-nk3yj]

Plzz make video of this project in Java language

[@programming_hut]

This is similar tensorflow android | Deploy deep learning model android studio | Java | ml kzbin.info/www/bejne/qqq2mI1np6qVjZY

[@IqraZahid-nk3yj]

Sir plzzz I have to submit my project in 2 days can you help me..... Make video in java language

[@programming_hut]

I have already uploaded that Deploy Ml model on android | android project java | android studio | cat dog app kzbin.info/www/bejne/l4C8pZehr76gd7s

[@surajbogati9818]

does this work for our own model as well?

[@programming_hut]

You can follow steps accordingly

[@dikshasalunke1535]

@@programming_hut sir i want the output in voive format... so where I put TTS code plz help

[@haritmengar]

bro can you give me this source code download link

[@syedshorifulalamopu9470]

10.1

[@triciamoscare7475]

How to create a tflite file with metadata if i am going to use my own model?

[@programming_hut]

You tried official documentation ? www.tensorflow.org/lite/models/convert/metadata#model_information

[@rambabupalla959]

Please can any one provide entire code I am getting errors

[@sushantkandwal]

did you have the code ?

[@ravindratiwari2993]

very fast explained...disappointing

[@S0da0DA]

Can you make this, but with YOLOv8 TFLite?

[@vishalkumaryadav8192]

hello, did you find any with yolov8 tflite?