How to run deep networks on Android device#
See also
Introduction#
In this tutorial you’ll know how to run deep learning networks on Android device using OpenCV deep learning module. Tutorial was written for Android Studio 2022.2.1.
Requirements#
Download and install Android Studio from https://developer.android.com/studio.
Get the latest pre-built OpenCV for Android release from opencv/opencv and unpack it (for example,
opencv-4.X.Y-android-sdk.zip, minimum version 4.9 is required).Download the MobileNet-SSD object detection model in ONNX format and save it as
mobilenet.onnx. This single file contains both the network topology and its trained weights.
Create an empty Android Studio project and add OpenCV dependency#
Use Android Development with OpenCV tutorial to initialize your project and add OpenCV.
Make an app#
Our sample will takes pictures from a camera, forwards it into a deep network and receives a set of rectangles, class identifiers and confidence values in range [0, 1].
First of all, we need to add a necessary widget which displays processed frames. Modify
app/src/main/res/layout/activity_main.xml:
<?xml version="1.0" encoding="utf-8"?>
<FrameLayout xmlns:android="http://schemas.android.com/apk/res/android"
xmlns:app="http://schemas.android.com/apk/res-auto"
xmlns:tools="http://schemas.android.com/tools"
android:layout_width="match_parent"
android:layout_height="match_parent"
tools:context="org.opencv.samples.opencv_mobilenet.MainActivity">
<org.opencv.android.JavaCameraView
android:id="@+id/CameraView"
android:layout_width="match_parent"
android:layout_height="match_parent"
android:visibility="visible" />
</FrameLayout>
Modify
/app/src/main/AndroidManifest.xmlto enable full-screen mode, set up a correct screen orientation and allow to use a camera.
<?xml version="1.0" encoding="utf-8"?>
<manifest xmlns:android="http://schemas.android.com/apk/res/android">
<application
android:label="@string/app_name">
<activity
android:exported="true"
android:name=".MainActivity"
android:screenOrientation="landscape"> <!--Screen orientation-->
<intent-filter>
<action android:name="android.intent.action.MAIN" />
<category android:name="android.intent.category.LAUNCHER" />
</intent-filter>
</activity>
</application>
<!--Allow to use a camera-->
<uses-permission android:name="android.permission.CAMERA"/>
<uses-feature android:name="android.hardware.camera" android:required="false"/>
<uses-feature android:name="android.hardware.camera.autofocus" android:required="false"/>
<uses-feature android:name="android.hardware.camera.front" android:required="false"/>
<uses-feature android:name="android.hardware.camera.front.autofocus" android:required="false"/>
</manifest>
Replace content of
app/src/main/java/com/example/myapplication/MainActivity.javaand set a custom package name if necessary:
package com.example.myapplication;
import android.content.Context;
import android.content.res.AssetManager;
import android.os.Bundle;
import android.util.Log;
import android.widget.Toast;
import org.opencv.android.CameraActivity;
import org.opencv.android.CameraBridgeViewBase;
import org.opencv.android.CameraBridgeViewBase.CvCameraViewFrame;
import org.opencv.android.CameraBridgeViewBase.CvCameraViewListener2;
import org.opencv.android.OpenCVLoader;
import org.opencv.core.Core;
import org.opencv.core.Mat;
import org.opencv.core.MatOfByte;
import org.opencv.core.Point;
import org.opencv.core.Scalar;
import org.opencv.core.Size;
import org.opencv.dnn.Net;
import org.opencv.dnn.Dnn;
import org.opencv.imgproc.Imgproc;
import java.io.InputStream;
import java.io.IOException;
import java.util.Collections;
import java.util.List;
public class MainActivity extends CameraActivity implements CvCameraViewListener2 {
@Override
public void onResume() {
super.onResume();
if (mOpenCvCameraView != null)
mOpenCvCameraView.enableView();
}
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
if (OpenCVLoader.initLocal()) {
Log.i(TAG, "OpenCV loaded successfully");
} else {
Log.e(TAG, "OpenCV initialization failed!");
(Toast.makeText(this, "OpenCV initialization failed!", Toast.LENGTH_LONG)).show();
return;
}
// Load a single-file MobileNet-SSD ONNX model from res/raw/mobilenet.onnx.
// The resource is resolved by name at runtime so the sample still builds when the
// (optionally downloaded) model is absent; object detection is simply disabled then.
int modelResId = getResources().getIdentifier("mobilenet", "raw", getPackageName());
if (modelResId != 0)
mModelBuffer = loadFileFromResource(modelResId);
if (mModelBuffer == null) {
Log.e(TAG, "MobileNet ONNX model (res/raw/mobilenet.onnx) not found - object detection disabled");
} else {
net = Dnn.readNetFromONNX(mModelBuffer);
Log.i(TAG, "Network loaded successfully");
}
setContentView(R.layout.activity_main);
// Set up camera listener.
mOpenCvCameraView = (CameraBridgeViewBase)findViewById(R.id.CameraView);
mOpenCvCameraView.setVisibility(CameraBridgeViewBase.VISIBLE);
mOpenCvCameraView.setCvCameraViewListener(this);
}
@Override
public void onPause()
{
super.onPause();
if (mOpenCvCameraView != null)
mOpenCvCameraView.disableView();
}
@Override
protected List<? extends CameraBridgeViewBase> getCameraViewList() {
return Collections.singletonList(mOpenCvCameraView);
}
public void onDestroy() {
super.onDestroy();
if (mOpenCvCameraView != null)
mOpenCvCameraView.disableView();
if (mModelBuffer != null)
mModelBuffer.release();
}
// Load a network.
public void onCameraViewStarted(int width, int height) {
}
public Mat onCameraFrame(CvCameraViewFrame inputFrame) {
final int IN_WIDTH = 300;
final int IN_HEIGHT = 300;
final float WH_RATIO = (float)IN_WIDTH / IN_HEIGHT;
final double IN_SCALE_FACTOR = 0.007843;
final double MEAN_VAL = 127.5;
final double THRESHOLD = 0.2;
if (net == null) {
return inputFrame.rgba();
}
Log.d(TAG, "handle new frame!");
Mat frame = inputFrame.rgba();
Imgproc.cvtColor(frame, frame, Imgproc.COLOR_RGBA2RGB);
// Forward image through network.
Mat blob = Dnn.blobFromImage(frame, IN_SCALE_FACTOR,
new Size(IN_WIDTH, IN_HEIGHT),
new Scalar(MEAN_VAL, MEAN_VAL, MEAN_VAL), /*swapRB*/false, /*crop*/false);
net.setInput(blob);
Mat detections = net.forward();
int cols = frame.cols();
int rows = frame.rows();
detections = detections.reshape(1, (int)detections.total() / 7);
for (int i = 0; i < detections.rows(); ++i) {
double confidence = detections.get(i, 2)[0];
if (confidence > THRESHOLD) {
int classId = (int)detections.get(i, 1)[0];
int left = (int)(detections.get(i, 3)[0] * cols);
int top = (int)(detections.get(i, 4)[0] * rows);
int right = (int)(detections.get(i, 5)[0] * cols);
int bottom = (int)(detections.get(i, 6)[0] * rows);
// Draw rectangle around detected object.
Imgproc.rectangle(frame, new Point(left, top), new Point(right, bottom),
new Scalar(0, 255, 0));
String label = classNames[classId] + ": " + confidence;
int[] baseLine = new int[1];
Size labelSize = Imgproc.getTextSize(label, Imgproc.FONT_HERSHEY_SIMPLEX, 0.5, 1, baseLine);
// Draw background for label.
Imgproc.rectangle(frame, new Point(left, top - labelSize.height),
new Point(left + labelSize.width, top + baseLine[0]),
new Scalar(255, 255, 255), Imgproc.FILLED);
// Write class name and confidence.
Imgproc.putText(frame, label, new Point(left, top),
Imgproc.FONT_HERSHEY_SIMPLEX, 0.5, new Scalar(0, 0, 0));
}
}
return frame;
}
public void onCameraViewStopped() {}
private MatOfByte loadFileFromResource(int id) {
byte[] buffer;
try {
// load cascade file from application resources
InputStream is = getResources().openRawResource(id);
int size = is.available();
buffer = new byte[size];
int bytesRead = is.read(buffer);
is.close();
} catch (IOException e) {
e.printStackTrace();
Log.e(TAG, "Failed to ONNX model from resources! Exception thrown: " + e);
(Toast.makeText(this, "Failed to ONNX model from resources!", Toast.LENGTH_LONG)).show();
return null;
}
return new MatOfByte(buffer);
}
private static final String TAG = "OpenCV-MobileNet";
private static final String[] classNames = {"background",
"aeroplane", "bicycle", "bird", "boat",
"bottle", "bus", "car", "cat", "chair",
"cow", "diningtable", "dog", "horse",
"motorbike", "person", "pottedplant",
"sheep", "sofa", "train", "tvmonitor"};
private MatOfByte mModelBuffer;
private Net net;
private CameraBridgeViewBase mOpenCvCameraView;
}
Put the downloaded
mobilenet.onnxintoapp/src/main/res/rawfolder. OpenCV DNN model is mainly designed to load ML and DNN models from file. Modern Android does not allow it without extra permissions, but provides Java API to load bytes from resources. The sample uses alternative DNN API that initializes a model from in-memory buffer rather than a file. The following function reads model file from resources and converts it toMatOfBytes(analog ofstd::vector<char>in C++ world) object suitable for OpenCV Java API:
private MatOfByte loadFileFromResource(int id) {
byte[] buffer;
try {
// load cascade file from application resources
InputStream is = getResources().openRawResource(id);
int size = is.available();
buffer = new byte[size];
int bytesRead = is.read(buffer);
is.close();
} catch (IOException e) {
e.printStackTrace();
Log.e(TAG, "Failed to ONNX model from resources! Exception thrown: " + e);
(Toast.makeText(this, "Failed to ONNX model from resources!", Toast.LENGTH_LONG)).show();
return null;
}
return new MatOfByte(buffer);
}
And then the network initialization is done with the following lines:
// Load a single-file MobileNet-SSD ONNX model from res/raw/mobilenet.onnx.
// The resource is resolved by name at runtime so the sample still builds when the
// (optionally downloaded) model is absent; object detection is simply disabled then.
int modelResId = getResources().getIdentifier("mobilenet", "raw", getPackageName());
if (modelResId != 0)
mModelBuffer = loadFileFromResource(modelResId);
if (mModelBuffer == null) {
Log.e(TAG, "MobileNet ONNX model (res/raw/mobilenet.onnx) not found - object detection disabled");
} else {
net = Dnn.readNetFromONNX(mModelBuffer);
Log.i(TAG, "Network loaded successfully");
}
See also Android documentation on resources
Take a look how DNN model input is prepared and inference result is interpreted:
Mat blob = Dnn.blobFromImage(frame, IN_SCALE_FACTOR,
new Size(IN_WIDTH, IN_HEIGHT),
new Scalar(MEAN_VAL, MEAN_VAL, MEAN_VAL), /*swapRB*/false, /*crop*/false);
net.setInput(blob);
Mat detections = net.forward();
int cols = frame.cols();
int rows = frame.rows();
detections = detections.reshape(1, (int)detections.total() / 7);
for (int i = 0; i < detections.rows(); ++i) {
double confidence = detections.get(i, 2)[0];
if (confidence > THRESHOLD) {
int classId = (int)detections.get(i, 1)[0];
int left = (int)(detections.get(i, 3)[0] * cols);
int top = (int)(detections.get(i, 4)[0] * rows);
int right = (int)(detections.get(i, 5)[0] * cols);
int bottom = (int)(detections.get(i, 6)[0] * rows);
// Draw rectangle around detected object.
Imgproc.rectangle(frame, new Point(left, top), new Point(right, bottom),
new Scalar(0, 255, 0));
String label = classNames[classId] + ": " + confidence;
int[] baseLine = new int[1];
Size labelSize = Imgproc.getTextSize(label, Imgproc.FONT_HERSHEY_SIMPLEX, 0.5, 1, baseLine);
// Draw background for label.
Imgproc.rectangle(frame, new Point(left, top - labelSize.height),
new Point(left + labelSize.width, top + baseLine[0]),
new Scalar(255, 255, 255), Imgproc.FILLED);
// Write class name and confidence.
Imgproc.putText(frame, label, new Point(left, top),
Imgproc.FONT_HERSHEY_SIMPLEX, 0.5, new Scalar(0, 0, 0));
}
}
Dnn.blobFromImage converts camera frame to neural network input tensor. Resize and statistical
normalization are applied. Each line of network output tensor contains information on one detected
object in the following order: confidence in range [0, 1], class id, left, top, right, bottom box
coordinates. All coordinates are in range [0, 1] and should be scaled to image size before rendering.
Launch an application and make a fun!
