TensorFlow Lite Object Detection FTP

Overview

This is a camera app that continuously detects the objects (bounding boxes and classes) in the frames seen by your device's back camera and send object image to FTP server, with the option to use a quantized MobileNet SSD, EfficientDet Lite 0, EfficientDet Lite1, or EfficientDet Lite2 model trained on the COCO dataset. These instructions walk you through building and running the demo on an Android device.

The model files are downloaded via Gradle scripts when you build and run the app. You don't need to do any steps to download TFLite models into the project explicitly.

This application should be run on a physical Android device.

Releases

Ver 2.2.0 changes:

Enhanced FTP action algorithm by adding switch

Ver 2.1.0 changes:

Developed FTP action on object detection

Ver 2.0.0 changes:

Added FTP action on object detection

Changed default values and selections in option to best

Downloaded tflite models to assets folder

Ver 1.0.0 changes:

Created repository

Build

Prerequisites

• This sample has been tested on Android IDE and Colab.

• A physical Android device with a minimum OS version of SDK 24 (Android 7.0 - Nougat) with developer mode enabled. The process of enabling developer mode may vary by device.

Building

Run below commands in Colab:

!wget https://dl.google.com/android/repository/commandlinetools-linux-9123335_latest.zip
!mkdir -p sdk
!unzip commandlinetools-linux-9123335_latest.zip -d sdk
!yes | ./sdk/cmdline-tools/bin/sdkmanager --sdk_root=/content/sdk "tools"
!git clone https://github.com/marzban2030/Camera-tflite-FTP
!chmod -c 755 /content/Camera-tflite-FTP/gradlew
!export ANDROID_HOME=/content/sdk && cd /content/Camera-tflite-FTP && ./gradlew assembleDebug

Then run:

from google.colab import files
files.download('Camera-tflite-FTP/app/build/outputs/apk/debug/app-debug.apk')

Usage

Change "person" in "ObjectDetectorHelper.kt" file according to tflite labels if you want to getting action on another object detection, Also change "ftphost", "ftpusername", "ftppassword", "/ftpdirectory" to be working.

... == "person") {
        ...
        ... ftpConnect("ftphost", "ftpusername", "ftppassword", 21)
        ...
        ... ftpUploadBitmap(bitmap, "/ftpdirectory" + ...

Models used

Downloading, extraction, and placing the models into the assets folder is managed automatically by the download.gradle file.