Skip to content

GitLab

"vscode:/vscode.git/clone" did not exist on "02519d50e3b5a950f9fb66654d004303d2f95a28"
Commit 4b214948 authored by zhiminzhang0830's avatar zhiminzhang0830
Browse files

Merge branch 'dygraph' of https://github.com/PaddlePaddle/PaddleOCR into new_branch

parents 917606b4 6e607a0f
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 358 additions and 573 deletions
deploy/android_demo/app/src/main/java/com/baidu/paddle/lite/demo/ocr/MainActivity.java
View file @ 4b214948
......@@ -13,6 +13,7 @@ import android.graphics.BitmapFactory;
import android.graphics.drawable.BitmapDrawable;
import android.media.ExifInterface;
import android.content.res.AssetManager;
import android.media.FaceDetector;
import android.net.Uri;
import android.os.Bundle;
import android.os.Environment;
......@@ -27,7 +28,9 @@ import android.view.Menu;
import android.view.MenuInflater;
import android.view.MenuItem;
import android.view.View;
import android.widget.CheckBox;
import android.widget.ImageView;
import android.widget.Spinner;
import android.widget.TextView;
import android.widget.Toast;
......@@ -68,23 +71,24 @@ public class MainActivity extends AppCompatActivity {
protected ImageView ivInputImage;
protected TextView tvOutputResult;
protected TextView tvInferenceTime;
protected CheckBox cbOpencl;
protected Spinner spRunMode;
// Model settings of object detection
// Model settings of ocr
protected String modelPath = "";
protected String labelPath = "";
protected String imagePath = "";
protected int cpuThreadNum = 1;
protected String cpuPowerMode = "";
protected String inputColorFormat = "";
protected long[] inputShape = new long[]{};
protected float[] inputMean = new float[]{};
protected float[] inputStd = new float[]{};
protected int detLongSize = 960;
protected float scoreThreshold = 0.1f;
private String currentPhotoPath;
private AssetManager assetManager =null;
private AssetManager assetManager = null;
protected Predictor predictor = new Predictor();
private Bitmap cur_predict_image = null;
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
......@@ -98,10 +102,12 @@ public class MainActivity extends AppCompatActivity {
// Setup the UI components
tvInputSetting = findViewById(R.id.tv_input_setting);
cbOpencl = findViewById(R.id.cb_opencl);
tvStatus = findViewById(R.id.tv_model_img_status);
ivInputImage = findViewById(R.id.iv_input_image);
tvInferenceTime = findViewById(R.id.tv_inference_time);
tvOutputResult = findViewById(R.id.tv_output_result);
spRunMode = findViewById(R.id.sp_run_mode);
tvInputSetting.setMovementMethod(ScrollingMovementMethod.getInstance());
tvOutputResult.setMovementMethod(ScrollingMovementMethod.getInstance());
......@@ -111,26 +117,26 @@ public class MainActivity extends AppCompatActivity {
public void handleMessage(Message msg) {
switch (msg.what) {
case RESPONSE_LOAD_MODEL_SUCCESSED:
if(pbLoadModel!=null && pbLoadModel.isShowing()){
if (pbLoadModel != null && pbLoadModel.isShowing()) {
pbLoadModel.dismiss();
}
onLoadModelSuccessed();
break;
case RESPONSE_LOAD_MODEL_FAILED:
if(pbLoadModel!=null && pbLoadModel.isShowing()){
if (pbLoadModel != null && pbLoadModel.isShowing()) {
pbLoadModel.dismiss();
}
Toast.makeText(MainActivity.this, "Load model failed!", Toast.LENGTH_SHORT).show();
onLoadModelFailed();
break;
case RESPONSE_RUN_MODEL_SUCCESSED:
if(pbRunModel!=null && pbRunModel.isShowing()){
if (pbRunModel != null && pbRunModel.isShowing()) {
pbRunModel.dismiss();
}
onRunModelSuccessed();
break;
case RESPONSE_RUN_MODEL_FAILED:
if(pbRunModel!=null && pbRunModel.isShowing()){
if (pbRunModel != null && pbRunModel.isShowing()) {
pbRunModel.dismiss();
}
Toast.makeText(MainActivity.this, "Run model failed!", Toast.LENGTH_SHORT).show();
......@@ -175,71 +181,47 @@ public class MainActivity extends AppCompatActivity {
super.onResume();
SharedPreferences sharedPreferences = PreferenceManager.getDefaultSharedPreferences(this);
boolean settingsChanged = false;
boolean model_settingsChanged = false;
String model_path = sharedPreferences.getString(getString(R.string.MODEL_PATH_KEY),
getString(R.string.MODEL_PATH_DEFAULT));
String label_path = sharedPreferences.getString(getString(R.string.LABEL_PATH_KEY),
getString(R.string.LABEL_PATH_DEFAULT));
String image_path = sharedPreferences.getString(getString(R.string.IMAGE_PATH_KEY),
getString(R.string.IMAGE_PATH_DEFAULT));
settingsChanged |= !model_path.equalsIgnoreCase(modelPath);
model_settingsChanged |= !model_path.equalsIgnoreCase(modelPath);
settingsChanged |= !label_path.equalsIgnoreCase(labelPath);
settingsChanged |= !image_path.equalsIgnoreCase(imagePath);
int cpu_thread_num = Integer.parseInt(sharedPreferences.getString(getString(R.string.CPU_THREAD_NUM_KEY),
getString(R.string.CPU_THREAD_NUM_DEFAULT)));
settingsChanged |= cpu_thread_num != cpuThreadNum;
model_settingsChanged |= cpu_thread_num != cpuThreadNum;
String cpu_power_mode =
sharedPreferences.getString(getString(R.string.CPU_POWER_MODE_KEY),
getString(R.string.CPU_POWER_MODE_DEFAULT));
settingsChanged |= !cpu_power_mode.equalsIgnoreCase(cpuPowerMode);
String input_color_format =
sharedPreferences.getString(getString(R.string.INPUT_COLOR_FORMAT_KEY),
getString(R.string.INPUT_COLOR_FORMAT_DEFAULT));
settingsChanged |= !input_color_format.equalsIgnoreCase(inputColorFormat);
long[] input_shape =
Utils.parseLongsFromString(sharedPreferences.getString(getString(R.string.INPUT_SHAPE_KEY),
getString(R.string.INPUT_SHAPE_DEFAULT)), ",");
float[] input_mean =
Utils.parseFloatsFromString(sharedPreferences.getString(getString(R.string.INPUT_MEAN_KEY),
getString(R.string.INPUT_MEAN_DEFAULT)), ",");
float[] input_std =
Utils.parseFloatsFromString(sharedPreferences.getString(getString(R.string.INPUT_STD_KEY)
, getString(R.string.INPUT_STD_DEFAULT)), ",");
settingsChanged |= input_shape.length != inputShape.length;
settingsChanged |= input_mean.length != inputMean.length;
settingsChanged |= input_std.length != inputStd.length;
if (!settingsChanged) {
for (int i = 0; i < input_shape.length; i++) {
settingsChanged |= input_shape[i] != inputShape[i];
}
for (int i = 0; i < input_mean.length; i++) {
settingsChanged |= input_mean[i] != inputMean[i];
}
for (int i = 0; i < input_std.length; i++) {
settingsChanged |= input_std[i] != inputStd[i];
}
}
model_settingsChanged |= !cpu_power_mode.equalsIgnoreCase(cpuPowerMode);
int det_long_size = Integer.parseInt(sharedPreferences.getString(getString(R.string.DET_LONG_SIZE_KEY),
getString(R.string.DET_LONG_SIZE_DEFAULT)));
settingsChanged |= det_long_size != detLongSize;
float score_threshold =
Float.parseFloat(sharedPreferences.getString(getString(R.string.SCORE_THRESHOLD_KEY),
getString(R.string.SCORE_THRESHOLD_DEFAULT)));
settingsChanged |= scoreThreshold != score_threshold;
if (settingsChanged) {
modelPath = model_path;
labelPath = label_path;
imagePath = image_path;
detLongSize = det_long_size;
scoreThreshold = score_threshold;
set_img();
}
if (model_settingsChanged) {
modelPath = model_path;
cpuThreadNum = cpu_thread_num;
cpuPowerMode = cpu_power_mode;
inputColorFormat = input_color_format;
inputShape = input_shape;
inputMean = input_mean;
inputStd = input_std;
scoreThreshold = score_threshold;
// Update UI
tvInputSetting.setText("Model: " + modelPath.substring(modelPath.lastIndexOf("/") + 1) + "\n" + "CPU" +
" Thread Num: " + Integer.toString(cpuThreadNum) + "\n" + "CPU Power Mode: " + cpuPowerMode);
tvInputSetting.setText("Model: " + modelPath.substring(modelPath.lastIndexOf("/") + 1) + "\nOPENCL: " + cbOpencl.isChecked() + "\nCPU Thread Num: " + cpuThreadNum + "\nCPU Power Mode: " + cpuPowerMode);
tvInputSetting.scrollTo(0, 0);
// Reload model if configure has been changed
// loadModel();
set_img();
loadModel();
}
}
......@@ -254,20 +236,28 @@ public class MainActivity extends AppCompatActivity {
}
public boolean onLoadModel() {
return predictor.init(MainActivity.this, modelPath, labelPath, cpuThreadNum,
if (predictor.isLoaded()) {
predictor.releaseModel();
}
return predictor.init(MainActivity.this, modelPath, labelPath, cbOpencl.isChecked() ? 1 : 0, cpuThreadNum,
cpuPowerMode,
inputColorFormat,
inputShape, inputMean,
inputStd, scoreThreshold);
detLongSize, scoreThreshold);
}
public boolean onRunModel() {
return predictor.isLoaded() && predictor.runModel();
String run_mode = spRunMode.getSelectedItem().toString();
int run_det = run_mode.contains("检测") ? 1 : 0;
int run_cls = run_mode.contains("分类") ? 1 : 0;
int run_rec = run_mode.contains("识别") ? 1 : 0;
return predictor.isLoaded() && predictor.runModel(run_det, run_cls, run_rec);
}
public void onLoadModelSuccessed() {
// Load test image from path and run model
tvInputSetting.setText("Model: " + modelPath.substring(modelPath.lastIndexOf("/") + 1) + "\nOPENCL: " + cbOpencl.isChecked() + "\nCPU Thread Num: " + cpuThreadNum + "\nCPU Power Mode: " + cpuPowerMode);
tvInputSetting.scrollTo(0, 0);
tvStatus.setText("STATUS: load model successed");
}
public void onLoadModelFailed() {
......@@ -290,20 +280,13 @@ public class MainActivity extends AppCompatActivity {
tvStatus.setText("STATUS: run model failed");
}
public void onImageChanged(Bitmap image) {
// Rerun model if users pick test image from gallery or camera
if (image != null && predictor.isLoaded()) {
predictor.setInputImage(image);
runModel();
}
}
public void set_img() {
// Load test image from path and run model
try {
assetManager= getAssets();
InputStream in=assetManager.open(imagePath);
Bitmap bmp=BitmapFactory.decodeStream(in);
assetManager = getAssets();
InputStream in = assetManager.open(imagePath);
Bitmap bmp = BitmapFactory.decodeStream(in);
cur_predict_image = bmp;
ivInputImage.setImageBitmap(bmp);
} catch (IOException e) {
Toast.makeText(MainActivity.this, "Load image failed!", Toast.LENGTH_SHORT).show();
......@@ -430,7 +413,7 @@ public class MainActivity extends AppCompatActivity {
Cursor cursor = managedQuery(uri, proj, null, null, null);
cursor.moveToFirst();
if (image != null) {
// onImageChanged(image);
cur_predict_image = image;
ivInputImage.setImageBitmap(image);
}
} catch (IOException e) {
......@@ -451,7 +434,7 @@ public class MainActivity extends AppCompatActivity {
Bitmap image = BitmapFactory.decodeFile(currentPhotoPath);
image = Utils.rotateBitmap(image, orientation);
if (image != null) {
// onImageChanged(image);
cur_predict_image = image;
ivInputImage.setImageBitmap(image);
}
} else {
......@@ -464,28 +447,28 @@ public class MainActivity extends AppCompatActivity {
}
}
public void btn_load_model_click(View view) {
if (predictor.isLoaded()){
tvStatus.setText("STATUS: model has been loaded");
}else{
tvStatus.setText("STATUS: load model ......");
loadModel();
}
public void btn_reset_img_click(View view) {
ivInputImage.setImageBitmap(cur_predict_image);
}
public void cb_opencl_click(View view) {
tvStatus.setText("STATUS: load model ......");
loadModel();
}
public void btn_run_model_click(View view) {
Bitmap image =((BitmapDrawable)ivInputImage.getDrawable()).getBitmap();
if(image == null) {
Bitmap image = ((BitmapDrawable) ivInputImage.getDrawable()).getBitmap();
if (image == null) {
tvStatus.setText("STATUS: image is not exists");
}
else if (!predictor.isLoaded()){
} else if (!predictor.isLoaded()) {
tvStatus.setText("STATUS: model is not loaded");
}else{
} else {
tvStatus.setText("STATUS: run model ...... ");
predictor.setInputImage(image);
runModel();
}
}
public void btn_choice_img_click(View view) {
if (requestAllPermissions()) {
openGallery();
......@@ -506,4 +489,32 @@ public class MainActivity extends AppCompatActivity {
worker.quit();
super.onDestroy();
}
public int get_run_mode() {
String run_mode = spRunMode.getSelectedItem().toString();
int mode;
switch (run_mode) {
case "检测+分类+识别":
mode = 1;
break;
case "检测+识别":
mode = 2;
break;
case "识别+分类":
mode = 3;
break;
case "检测":
mode = 4;
break;
case "识别":
mode = 5;
break;
case "分类":
mode = 6;
break;
default:
mode = 1;
}
return mode;
}
}
deploy/android_demo/app/src/main/java/com/baidu/paddle/lite/demo/ocr/MiniActivity.java deleted 100644 → 0
View file @ 917606b4
package com.baidu.paddle.lite.demo.ocr;
import android.graphics.Bitmap;
import android.graphics.BitmapFactory;
import android.os.Build;
import android.os.Bundle;
import android.os.Handler;
import android.os.HandlerThread;
import android.os.Message;
import android.util.Log;
import android.view.View;
import android.widget.Button;
import android.widget.ImageView;
import android.widget.TextView;
import android.widget.Toast;
import androidx.appcompat.app.AppCompatActivity;
import java.io.IOException;
import java.io.InputStream;
public class MiniActivity extends AppCompatActivity {
public static final int REQUEST_LOAD_MODEL = 0;
public static final int REQUEST_RUN_MODEL = 1;
public static final int REQUEST_UNLOAD_MODEL = 2;
public static final int RESPONSE_LOAD_MODEL_SUCCESSED = 0;
public static final int RESPONSE_LOAD_MODEL_FAILED = 1;
public static final int RESPONSE_RUN_MODEL_SUCCESSED = 2;
public static final int RESPONSE_RUN_MODEL_FAILED = 3;
private static final String TAG = "MiniActivity";
protected Handler receiver = null; // Receive messages from worker thread
protected Handler sender = null; // Send command to worker thread
protected HandlerThread worker = null; // Worker thread to load&run model
protected volatile Predictor predictor = null;
private String assetModelDirPath = "models/ocr_v2_for_cpu";
private String assetlabelFilePath = "labels/ppocr_keys_v1.txt";
private Button button;
private ImageView imageView; // image result
private TextView textView; // text result
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.activity_mini);
Log.i(TAG, "SHOW in Logcat");
// Prepare the worker thread for mode loading and inference
worker = new HandlerThread("Predictor Worker");
worker.start();
sender = new Handler(worker.getLooper()) {
public void handleMessage(Message msg) {
switch (msg.what) {
case REQUEST_LOAD_MODEL:
// Load model and reload test image
if (!onLoadModel()) {
runOnUiThread(new Runnable() {
@Override
public void run() {
Toast.makeText(MiniActivity.this, "Load model failed!", Toast.LENGTH_SHORT).show();
}
});
}
break;
case REQUEST_RUN_MODEL:
// Run model if model is loaded
final boolean isSuccessed = onRunModel();
runOnUiThread(new Runnable() {
@Override
public void run() {
if (isSuccessed){
onRunModelSuccessed();
}else{
Toast.makeText(MiniActivity.this, "Run model failed!", Toast.LENGTH_SHORT).show();
}
}
});
break;
}
}
};
sender.sendEmptyMessage(REQUEST_LOAD_MODEL); // corresponding to REQUEST_LOAD_MODEL， to call onLoadModel()
imageView = findViewById(R.id.imageView);
textView = findViewById(R.id.sample_text);
button = findViewById(R.id.button);
button.setOnClickListener(new View.OnClickListener() {
@Override
public void onClick(View v) {
sender.sendEmptyMessage(REQUEST_RUN_MODEL);
}
});
}
@Override
protected void onDestroy() {
onUnloadModel();
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.JELLY_BEAN_MR2) {
worker.quitSafely();
} else {
worker.quit();
}
super.onDestroy();
}
/**
* call in onCreate, model init
*
* @return
*/
private boolean onLoadModel() {
if (predictor == null) {
predictor = new Predictor();
}
return predictor.init(this, assetModelDirPath, assetlabelFilePath);
}
/**
* init engine
* call in onCreate
*
* @return
*/
private boolean onRunModel() {
try {
String assetImagePath = "images/0.jpg";
InputStream imageStream = getAssets().open(assetImagePath);
Bitmap image = BitmapFactory.decodeStream(imageStream);
// Input is Bitmap
predictor.setInputImage(image);
return predictor.isLoaded() && predictor.runModel();
} catch (IOException e) {
e.printStackTrace();
return false;
}
}
private void onRunModelSuccessed() {
Log.i(TAG, "onRunModelSuccessed");
textView.setText(predictor.outputResult);
imageView.setImageBitmap(predictor.outputImage);
}
private void onUnloadModel() {
if (predictor != null) {
predictor.releaseModel();
}
}
}
deploy/android_demo/app/src/main/java/com/baidu/paddle/lite/demo/ocr/OCRPredictorNative.java
View file @ 4b214948
......@@ -29,22 +29,22 @@ public class OCRPredictorNative {
public OCRPredictorNative(Config config) {
this.config = config;
loadLibrary();
nativePointer = init(config.detModelFilename, config.recModelFilename,config.clsModelFilename,
nativePointer = init(config.detModelFilename, config.recModelFilename, config.clsModelFilename, config.useOpencl,
config.cpuThreadNum, config.cpuPower);
Log.i("OCRPredictorNative", "load success " + nativePointer);
}
public ArrayList<OcrResultModel> runImage(float[] inputData, int width, int height, int channels, Bitmap originalImage) {
Log.i("OCRPredictorNative", "begin to run image " + inputData.length + " " + width + " " + height);
float[] dims = new float[]{1, channels, height, width};
float[] rawResults = forward(nativePointer, inputData, dims, originalImage);
public ArrayList<OcrResultModel> runImage(Bitmap originalImage, int max_size_len, int run_det, int run_cls, int run_rec) {
Log.i("OCRPredictorNative", "begin to run image ");
float[] rawResults = forward(nativePointer, originalImage, max_size_len, run_det, run_cls, run_rec);
ArrayList<OcrResultModel> results = postprocess(rawResults);
return results;
}
public static class Config {
public int useOpencl;
public int cpuThreadNum;
public String cpuPower;
public String detModelFilename;
......@@ -53,16 +53,16 @@ public class OCRPredictorNative {
}
public void destory(){
public void destory() {
if (nativePointer > 0) {
release(nativePointer);
nativePointer = 0;
}
}
protected native long init(String detModelPath, String recModelPath,String clsModelPath, int threadNum, String cpuMode);
protected native long init(String detModelPath, String recModelPath, String clsModelPath, int useOpencl, int threadNum, String cpuMode);
protected native float[] forward(long pointer, float[] buf, float[] ddims, Bitmap originalImage);
protected native float[] forward(long pointer, Bitmap originalImage,int max_size_len, int run_det, int run_cls, int run_rec);
protected native void release(long pointer);
......@@ -73,9 +73,9 @@ public class OCRPredictorNative {
while (begin < raw.length) {
int point_num = Math.round(raw[begin]);
int word_num = Math.round(raw[begin + 1]);
OcrResultModel model = parse(raw, begin + 2, point_num, word_num);
begin += 2 + 1 + point_num * 2 + word_num;
results.add(model);
OcrResultModel res = parse(raw, begin + 2, point_num, word_num);
begin += 2 + 1 + point_num * 2 + word_num + 2;
results.add(res);
}
return results;
......@@ -83,19 +83,22 @@ public class OCRPredictorNative {
private OcrResultModel parse(float[] raw, int begin, int pointNum, int wordNum) {
int current = begin;
OcrResultModel model = new OcrResultModel();
model.setConfidence(raw[current]);
OcrResultModel res = new OcrResultModel();
res.setConfidence(raw[current]);
current++;
for (int i = 0; i < pointNum; i++) {
model.addPoints(Math.round(raw[current + i * 2]), Math.round(raw[current + i * 2 + 1]));
res.addPoints(Math.round(raw[current + i * 2]), Math.round(raw[current + i * 2 + 1]));
}
current += (pointNum * 2);
for (int i = 0; i < wordNum; i++) {
int index = Math.round(raw[current + i]);
model.addWordIndex(index);
res.addWordIndex(index);
}
current += wordNum;
res.setClsIdx(raw[current]);
res.setClsConfidence(raw[current + 1]);
Log.i("OCRPredictorNative", "word finished " + wordNum);
return model;
return res;
}
......
deploy/android_demo/app/src/main/java/com/baidu/paddle/lite/demo/ocr/OcrResultModel.java
View file @ 4b214948
......@@ -10,6 +10,9 @@ public class OcrResultModel {
private List<Integer> wordIndex;
private String label;
private float confidence;
private float cls_idx;
private String cls_label;
private float cls_confidence;
public OcrResultModel() {
super();
......@@ -49,4 +52,28 @@ public class OcrResultModel {
public void setConfidence(float confidence) {
this.confidence = confidence;
}
public float getClsIdx() {
return cls_idx;
}
public void setClsIdx(float idx) {
this.cls_idx = idx;
}
public String getClsLabel() {
return cls_label;
}
public void setClsLabel(String label) {
this.cls_label = label;
}
public float getClsConfidence() {
return cls_confidence;
}
public void setClsConfidence(float confidence) {
this.cls_confidence = confidence;
}
}
deploy/android_demo/app/src/main/java/com/baidu/paddle/lite/demo/ocr/Predictor.java
View file @ 4b214948
......@@ -31,23 +31,19 @@ public class Predictor {
protected float inferenceTime = 0;
// Only for object detection
protected Vector<String> wordLabels = new Vector<String>();
protected String inputColorFormat = "BGR";
protected long[] inputShape = new long[]{1, 3, 960};
protected float[] inputMean = new float[]{0.485f, 0.456f, 0.406f};
protected float[] inputStd = new float[]{1.0f / 0.229f, 1.0f / 0.224f, 1.0f / 0.225f};
protected int detLongSize = 960;
protected float scoreThreshold = 0.1f;
protected Bitmap inputImage = null;
protected Bitmap outputImage = null;
protected volatile String outputResult = "";
protected float preprocessTime = 0;
protected float postprocessTime = 0;
public Predictor() {
}
public boolean init(Context appCtx, String modelPath, String labelPath) {
isLoaded = loadModel(appCtx, modelPath, cpuThreadNum, cpuPowerMode);
public boolean init(Context appCtx, String modelPath, String labelPath, int useOpencl, int cpuThreadNum, String cpuPowerMode) {
isLoaded = loadModel(appCtx, modelPath, useOpencl, cpuThreadNum, cpuPowerMode);
if (!isLoaded) {
return false;
}
......@@ -56,49 +52,18 @@ public class Predictor {
}
public boolean init(Context appCtx, String modelPath, String labelPath, int cpuThreadNum, String cpuPowerMode,
String inputColorFormat,
long[] inputShape, float[] inputMean,
float[] inputStd, float scoreThreshold) {
if (inputShape.length != 3) {
Log.e(TAG, "Size of input shape should be: 3");
return false;
}
if (inputMean.length != inputShape[1]) {
Log.e(TAG, "Size of input mean should be: " + Long.toString(inputShape[1]));
return false;
}
if (inputStd.length != inputShape[1]) {
Log.e(TAG, "Size of input std should be: " + Long.toString(inputShape[1]));
return false;
}
if (inputShape[0] != 1) {
Log.e(TAG, "Only one batch is supported in the image classification demo, you can use any batch size in " +
"your Apps!");
return false;
}
if (inputShape[1] != 1 && inputShape[1] != 3) {
Log.e(TAG, "Only one/three channels are supported in the image classification demo, you can use any " +
"channel size in your Apps!");
return false;
}
if (!inputColorFormat.equalsIgnoreCase("BGR")) {
Log.e(TAG, "Only BGR color format is supported.");
return false;
}
boolean isLoaded = init(appCtx, modelPath, labelPath);
public boolean init(Context appCtx, String modelPath, String labelPath, int useOpencl, int cpuThreadNum, String cpuPowerMode,
int detLongSize, float scoreThreshold) {
boolean isLoaded = init(appCtx, modelPath, labelPath, useOpencl, cpuThreadNum, cpuPowerMode);
if (!isLoaded) {
return false;
}
this.inputColorFormat = inputColorFormat;
this.inputShape = inputShape;
this.inputMean = inputMean;
this.inputStd = inputStd;
this.detLongSize = detLongSize;
this.scoreThreshold = scoreThreshold;
return true;
}
protected boolean loadModel(Context appCtx, String modelPath, int cpuThreadNum, String cpuPowerMode) {
protected boolean loadModel(Context appCtx, String modelPath, int useOpencl, int cpuThreadNum, String cpuPowerMode) {
// Release model if exists
releaseModel();
......@@ -118,12 +83,13 @@ public class Predictor {
}
OCRPredictorNative.Config config = new OCRPredictorNative.Config();
config.useOpencl = useOpencl;
config.cpuThreadNum = cpuThreadNum;
config.detModelFilename = realPath + File.separator + "ch_ppocr_mobile_v2.0_det_opt.nb";
config.recModelFilename = realPath + File.separator + "ch_ppocr_mobile_v2.0_rec_opt.nb";
config.clsModelFilename = realPath + File.separator + "ch_ppocr_mobile_v2.0_cls_opt.nb";
Log.e("Predictor", "model path" + config.detModelFilename + " ; " + config.recModelFilename + ";" + config.clsModelFilename);
config.cpuPower = cpuPowerMode;
config.detModelFilename = realPath + File.separator + "det_db.nb";
config.recModelFilename = realPath + File.separator + "rec_crnn.nb";
config.clsModelFilename = realPath + File.separator + "cls.nb";
Log.i("Predictor", "model path" + config.detModelFilename + " ; " + config.recModelFilename + ";" + config.clsModelFilename);
paddlePredictor = new OCRPredictorNative(config);
this.cpuThreadNum = cpuThreadNum;
......@@ -170,82 +136,29 @@ public class Predictor {
}
public boolean runModel() {
public boolean runModel(int run_det, int run_cls, int run_rec) {
if (inputImage == null || !isLoaded()) {
return false;
}
// Pre-process image, and feed input tensor with pre-processed data
Bitmap scaleImage = Utils.resizeWithStep(inputImage, Long.valueOf(inputShape[2]).intValue(), 32);
Date start = new Date();
int channels = (int) inputShape[1];
int width = scaleImage.getWidth();
int height = scaleImage.getHeight();
float[] inputData = new float[channels * width * height];
if (channels == 3) {
int[] channelIdx = null;
if (inputColorFormat.equalsIgnoreCase("RGB")) {
channelIdx = new int[]{0, 1, 2};
} else if (inputColorFormat.equalsIgnoreCase("BGR")) {
channelIdx = new int[]{2, 1, 0};
} else {
Log.i(TAG, "Unknown color format " + inputColorFormat + ", only RGB and BGR color format is " +
"supported!");
return false;
}
int[] channelStride = new int[]{width * height, width * height * 2};
int[] pixels=new int[width*height];
scaleImage.getPixels(pixels,0,scaleImage.getWidth(),0,0,scaleImage.getWidth(),scaleImage.getHeight());
for (int i = 0; i < pixels.length; i++) {
int color = pixels[i];
float[] rgb = new float[]{(float) red(color) / 255.0f, (float) green(color) / 255.0f,
(float) blue(color) / 255.0f};
inputData[i] = (rgb[channelIdx[0]] - inputMean[0]) / inputStd[0];
inputData[i + channelStride[0]] = (rgb[channelIdx[1]] - inputMean[1]) / inputStd[1];
inputData[i+ channelStride[1]] = (rgb[channelIdx[2]] - inputMean[2]) / inputStd[2];
}
} else if (channels == 1) {
int[] pixels=new int[width*height];
scaleImage.getPixels(pixels,0,scaleImage.getWidth(),0,0,scaleImage.getWidth(),scaleImage.getHeight());
for (int i = 0; i < pixels.length; i++) {
int color = pixels[i];
float gray = (float) (red(color) + green(color) + blue(color)) / 3.0f / 255.0f;
inputData[i] = (gray - inputMean[0]) / inputStd[0];
}
} else {
Log.i(TAG, "Unsupported channel size " + Integer.toString(channels) + ", only channel 1 and 3 is " +
"supported!");
return false;
}
float[] pixels = inputData;
Log.i(TAG, "pixels " + pixels[0] + " " + pixels[1] + " " + pixels[2] + " " + pixels[3]
+ " " + pixels[pixels.length / 2] + " " + pixels[pixels.length / 2 + 1] + " " + pixels[pixels.length - 2] + " " + pixels[pixels.length - 1]);
Date end = new Date();
preprocessTime = (float) (end.getTime() - start.getTime());
// Warm up
for (int i = 0; i < warmupIterNum; i++) {
paddlePredictor.runImage(inputData, width, height, channels, inputImage);
paddlePredictor.runImage(inputImage, detLongSize, run_det, run_cls, run_rec);
}
warmupIterNum = 0; // do not need warm
// Run inference
start = new Date();
ArrayList<OcrResultModel> results = paddlePredictor.runImage(inputData, width, height, channels, inputImage);
end = new Date();
Date start = new Date();
ArrayList<OcrResultModel> results = paddlePredictor.runImage(inputImage, detLongSize, run_det, run_cls, run_rec);
Date end = new Date();
inferenceTime = (end.getTime() - start.getTime()) / (float) inferIterNum;
results = postprocess(results);
Log.i(TAG, "[stat] Preprocess Time: " + preprocessTime
+ " ; Inference Time: " + inferenceTime + " ;Box Size " + results.size());
Log.i(TAG, "[stat] Inference Time: " + inferenceTime + " ;Box Size " + results.size());
drawResults(results);
return true;
}
public boolean isLoaded() {
return paddlePredictor != null && isLoaded;
}
......@@ -282,10 +195,6 @@ public class Predictor {
return outputResult;
}
public float preprocessTime() {
return preprocessTime;
}
public float postprocessTime() {
return postprocessTime;
}
......@@ -310,6 +219,7 @@ public class Predictor {
}
}
r.setLabel(word.toString());
r.setClsLabel(r.getClsIdx() == 1 ? "180" : "0");
}
return results;
}
......@@ -319,14 +229,22 @@ public class Predictor {
for (int i = 0; i < results.size(); i++) {
OcrResultModel result = results.get(i);
StringBuilder sb = new StringBuilder("");
sb.append(result.getLabel());
sb.append(" ").append(result.getConfidence());
sb.append("; Points: ");
for (Point p : result.getPoints()) {
sb.append("(").append(p.x).append(",").append(p.y).append(") ");
if(result.getPoints().size()>0){
sb.append("Det: ");
for (Point p : result.getPoints()) {
sb.append("(").append(p.x).append(",").append(p.y).append(") ");
}
}
if(result.getLabel().length() > 0){
sb.append("\n Rec: ").append(result.getLabel());
sb.append(",").append(result.getConfidence());
}
if(result.getClsIdx()!=-1){
sb.append(" Cls: ").append(result.getClsLabel());
sb.append(",").append(result.getClsConfidence());
}
Log.i(TAG, sb.toString()); // show LOG in Logcat panel
outputResultSb.append(i + 1).append(": ").append(result.getLabel()).append("\n");
outputResultSb.append(i + 1).append(": ").append(sb.toString()).append("\n");
}
outputResult = outputResultSb.toString();
outputImage = inputImage;
......@@ -344,6 +262,9 @@ public class Predictor {
for (OcrResultModel result : results) {
Path path = new Path();
List<Point> points = result.getPoints();
if(points.size()==0){
continue;
}
path.moveTo(points.get(0).x, points.get(0).y);
for (int i = points.size() - 1; i >= 0; i--) {
Point p = points.get(i);
......
deploy/android_demo/app/src/main/java/com/baidu/paddle/lite/demo/ocr/SettingsActivity.java
View file @ 4b214948
......@@ -20,16 +20,13 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
ListPreference etImagePath = null;
ListPreference lpCPUThreadNum = null;
ListPreference lpCPUPowerMode = null;
ListPreference lpInputColorFormat = null;
EditTextPreference etInputShape = null;
EditTextPreference etInputMean = null;
EditTextPreference etInputStd = null;
EditTextPreference etDetLongSize = null;
EditTextPreference etScoreThreshold = null;
List<String> preInstalledModelPaths = null;
List<String> preInstalledLabelPaths = null;
List<String> preInstalledImagePaths = null;
List<String> preInstalledInputShapes = null;
List<String> preInstalledDetLongSizes = null;
List<String> preInstalledCPUThreadNums = null;
List<String> preInstalledCPUPowerModes = null;
List<String> preInstalledInputColorFormats = null;
......@@ -50,7 +47,7 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
preInstalledModelPaths = new ArrayList<String>();
preInstalledLabelPaths = new ArrayList<String>();
preInstalledImagePaths = new ArrayList<String>();
preInstalledInputShapes = new ArrayList<String>();
preInstalledDetLongSizes = new ArrayList<String>();
preInstalledCPUThreadNums = new ArrayList<String>();
preInstalledCPUPowerModes = new ArrayList<String>();
preInstalledInputColorFormats = new ArrayList<String>();
......@@ -63,10 +60,7 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
preInstalledImagePaths.add(getString(R.string.IMAGE_PATH_DEFAULT));
preInstalledCPUThreadNums.add(getString(R.string.CPU_THREAD_NUM_DEFAULT));
preInstalledCPUPowerModes.add(getString(R.string.CPU_POWER_MODE_DEFAULT));
preInstalledInputColorFormats.add(getString(R.string.INPUT_COLOR_FORMAT_DEFAULT));
preInstalledInputShapes.add(getString(R.string.INPUT_SHAPE_DEFAULT));
preInstalledInputMeans.add(getString(R.string.INPUT_MEAN_DEFAULT));
preInstalledInputStds.add(getString(R.string.INPUT_STD_DEFAULT));
preInstalledDetLongSizes.add(getString(R.string.DET_LONG_SIZE_DEFAULT));
preInstalledScoreThresholds.add(getString(R.string.SCORE_THRESHOLD_DEFAULT));
// Setup UI components
......@@ -89,11 +83,7 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
(ListPreference) findPreference(getString(R.string.CPU_THREAD_NUM_KEY));
lpCPUPowerMode =
(ListPreference) findPreference(getString(R.string.CPU_POWER_MODE_KEY));
lpInputColorFormat =
(ListPreference) findPreference(getString(R.string.INPUT_COLOR_FORMAT_KEY));
etInputShape = (EditTextPreference) findPreference(getString(R.string.INPUT_SHAPE_KEY));
etInputMean = (EditTextPreference) findPreference(getString(R.string.INPUT_MEAN_KEY));
etInputStd = (EditTextPreference) findPreference(getString(R.string.INPUT_STD_KEY));
etDetLongSize = (EditTextPreference) findPreference(getString(R.string.DET_LONG_SIZE_KEY));
etScoreThreshold = (EditTextPreference) findPreference(getString(R.string.SCORE_THRESHOLD_KEY));
}
......@@ -112,11 +102,7 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
editor.putString(getString(R.string.IMAGE_PATH_KEY), preInstalledImagePaths.get(modelIdx));
editor.putString(getString(R.string.CPU_THREAD_NUM_KEY), preInstalledCPUThreadNums.get(modelIdx));
editor.putString(getString(R.string.CPU_POWER_MODE_KEY), preInstalledCPUPowerModes.get(modelIdx));
editor.putString(getString(R.string.INPUT_COLOR_FORMAT_KEY),
preInstalledInputColorFormats.get(modelIdx));
editor.putString(getString(R.string.INPUT_SHAPE_KEY), preInstalledInputShapes.get(modelIdx));
editor.putString(getString(R.string.INPUT_MEAN_KEY), preInstalledInputMeans.get(modelIdx));
editor.putString(getString(R.string.INPUT_STD_KEY), preInstalledInputStds.get(modelIdx));
editor.putString(getString(R.string.DET_LONG_SIZE_KEY), preInstalledDetLongSizes.get(modelIdx));
editor.putString(getString(R.string.SCORE_THRESHOLD_KEY),
preInstalledScoreThresholds.get(modelIdx));
editor.apply();
......@@ -129,10 +115,7 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
etImagePath.setEnabled(enableCustomSettings);
lpCPUThreadNum.setEnabled(enableCustomSettings);
lpCPUPowerMode.setEnabled(enableCustomSettings);
lpInputColorFormat.setEnabled(enableCustomSettings);
etInputShape.setEnabled(enableCustomSettings);
etInputMean.setEnabled(enableCustomSettings);
etInputStd.setEnabled(enableCustomSettings);
etDetLongSize.setEnabled(enableCustomSettings);
etScoreThreshold.setEnabled(enableCustomSettings);
modelPath = sharedPreferences.getString(getString(R.string.MODEL_PATH_KEY),
getString(R.string.MODEL_PATH_DEFAULT));
......@@ -144,14 +127,8 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
getString(R.string.CPU_THREAD_NUM_DEFAULT));
String cpuPowerMode = sharedPreferences.getString(getString(R.string.CPU_POWER_MODE_KEY),
getString(R.string.CPU_POWER_MODE_DEFAULT));
String inputColorFormat = sharedPreferences.getString(getString(R.string.INPUT_COLOR_FORMAT_KEY),
getString(R.string.INPUT_COLOR_FORMAT_DEFAULT));
String inputShape = sharedPreferences.getString(getString(R.string.INPUT_SHAPE_KEY),
getString(R.string.INPUT_SHAPE_DEFAULT));
String inputMean = sharedPreferences.getString(getString(R.string.INPUT_MEAN_KEY),
getString(R.string.INPUT_MEAN_DEFAULT));
String inputStd = sharedPreferences.getString(getString(R.string.INPUT_STD_KEY),
getString(R.string.INPUT_STD_DEFAULT));
String detLongSize = sharedPreferences.getString(getString(R.string.DET_LONG_SIZE_KEY),
getString(R.string.DET_LONG_SIZE_DEFAULT));
String scoreThreshold = sharedPreferences.getString(getString(R.string.SCORE_THRESHOLD_KEY),
getString(R.string.SCORE_THRESHOLD_DEFAULT));
etModelPath.setSummary(modelPath);
......@@ -164,14 +141,8 @@ public class SettingsActivity extends AppCompatPreferenceActivity implements Sha
lpCPUThreadNum.setSummary(cpuThreadNum);
lpCPUPowerMode.setValue(cpuPowerMode);
lpCPUPowerMode.setSummary(cpuPowerMode);
lpInputColorFormat.setValue(inputColorFormat);
lpInputColorFormat.setSummary(inputColorFormat);
etInputShape.setSummary(inputShape);
etInputShape.setText(inputShape);
etInputMean.setSummary(inputMean);
etInputMean.setText(inputMean);
etInputStd.setSummary(inputStd);
etInputStd.setText(inputStd);
etDetLongSize.setSummary(detLongSize);
etDetLongSize.setText(detLongSize);
etScoreThreshold.setText(scoreThreshold);
etScoreThreshold.setSummary(scoreThreshold);
}
......
deploy/android_demo/app/src/main/res/layout/activity_main.xml
View file @ 4b214948
......@@ -23,13 +23,7 @@
android:layout_height="wrap_content"
android:orientation="horizontal">
<Button
android:id="@+id/btn_load_model"
android:layout_width="0dp"
android:layout_height="wrap_content"
android:layout_weight="1"
android:onClick="btn_load_model_click"
android:text="加载模型" />
<Button
android:id="@+id/btn_run_model"
android:layout_width="0dp"
......@@ -52,7 +46,45 @@
android:onClick="btn_choice_img_click"
android:text="选取图片" />
<Button
android:id="@+id/btn_reset_img"
android:layout_width="0dp"
android:layout_height="wrap_content"
android:layout_weight="1"
android:onClick="btn_reset_img_click"
android:text="清空绘图" />
</LinearLayout>
<LinearLayout
android:id="@+id/run_mode_layout"
android:layout_width="fill_parent"
android:layout_height="wrap_content"
android:orientation="horizontal">
<CheckBox
android:id="@+id/cb_opencl"
android:layout_width="0dp"
android:layout_weight="1"
android:layout_height="wrap_content"
android:text="开启OPENCL"
android:onClick="cb_opencl_click"
android:visibility="gone"/>
<TextView
android:layout_width="0dp"
android:layout_weight="0.5"
android:layout_height="wrap_content"
android:text="运行模式:"/>
<Spinner
android:id="@+id/sp_run_mode"
android:layout_width="0dp"
android:layout_weight="1.5"
android:layout_height="wrap_content"
android:entries="@array/run_Model"
/>
</LinearLayout>
<TextView
android:id="@+id/tv_input_setting"
android:layout_width="wrap_content"
......@@ -60,7 +92,7 @@
android:scrollbars="vertical"
android:layout_marginLeft="12dp"
android:layout_marginRight="12dp"
android:layout_marginTop="10dp"
android:layout_marginTop="5dp"
android:layout_marginBottom="5dp"
android:lineSpacingExtra="4dp"
android:singleLine="false"
......
deploy/android_demo/app/src/main/res/layout/activity_mini.xml deleted 100644 → 0
View file @ 917606b4
<?xml version="1.0" encoding="utf-8"?>
<!-- for MiniActivity Use Only -->
<androidx.constraintlayout.widget.ConstraintLayout 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"
app:layout_constraintLeft_toLeftOf="parent"
app:layout_constraintLeft_toRightOf="parent"
tools:context=".MainActivity">
<TextView
android:id="@+id/sample_text"
android:layout_width="0dp"
android:layout_height="wrap_content"
android:text="Hello World!"
app:layout_constraintLeft_toLeftOf="parent"
app:layout_constraintRight_toRightOf="parent"
app:layout_constraintTop_toBottomOf="@id/imageView"
android:scrollbars="vertical"
/>
<ImageView
android:id="@+id/imageView"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:paddingTop="20dp"
android:paddingBottom="20dp"
app:layout_constraintBottom_toTopOf="@id/imageView"
app:layout_constraintLeft_toLeftOf="parent"
app:layout_constraintRight_toRightOf="parent"
app:layout_constraintTop_toTopOf="parent"
tools:srcCompat="@tools:sample/avatars" />
<Button
android:id="@+id/button"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:layout_marginBottom="4dp"
android:text="Button"
app:layout_constraintBottom_toBottomOf="parent"
app:layout_constraintLeft_toLeftOf="parent"
app:layout_constraintRight_toRightOf="parent"
tools:layout_editor_absoluteX="161dp" />
</androidx.constraintlayout.widget.ConstraintLayout>
\ No newline at end of file
deploy/lite/readme.md
View file @ 4b214948
This diff is collapsed.
deploy/lite/readme_en.md
View file @ 4b214948
This diff is collapsed.
deploy/pdserving/README.md
View file @ 4b214948
This diff is collapsed.
deploy/slim/prune/README.md
View file @ 4b214948
......@@ -45,7 +45,7 @@ python3 setup.py install
'conv10_expand_weights': {0.1: 0.006509952684312718, 0.2: 0.01827734339798862, 0.3: 0.014528405644659832, 0.6: 0.06536008804270439, 0.8: 0.11798612250664964, 0.7: 0.12391408417493704, 0.4: 0.030615754498018757, 0.5: 0.047105205602406594}
'conv10_linear_weights': {0.1: 0.05113190831455035, 0.2: 0.07705573833558801, 0.3: 0.12096721757739311, 0.6: 0.5135061352930738, 0.8: 0.7908166677143281, 0.7: 0.7272187676899062, 0.4: 0.1819252083008504, 0.5: 0.3728054727792405}
}
加载敏感度文件后会返回一个字典，字典中的keys为网络模型参数模型的名字，values为一个字典，里面保存了相应网络层的裁剪敏感度信息。例如在例子中，conv10_expand_weights所对应的网络层在裁掉10%的卷积核后模型性能相较原模型会下降0.65%，详细信息可见[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim/blob/develop/docs/zh_cn/algo/algo.md#2-%E5%8D%B7%E7%A7%AF%E6%A0%B8%E5%89%AA%E8%A3%81%E5%8E%9F%E7%90%86)
加载敏感度文件后会返回一个字典，字典中的keys为网络模型参数模型的名字，values为一个字典，里面保存了相应网络层的裁剪敏感度信息。例如在例子中，conv10_expand_weights所对应的网络层在裁掉10%的卷积核后模型性能相较原模型会下降0.65%，详细信息可见[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim/blob/release/2.0-alpha/docs/zh_cn/algo/algo.md)
进入PaddleOCR根目录，通过以下命令对模型进行敏感度分析训练：
```bash
......
deploy/slim/prune/README_en.md
View file @ 4b214948
This diff is collapsed.
deploy/slim/quantization/README_en.md
View file @ 4b214948
This diff is collapsed.
doc/doc_ch/FAQ.md
View file @ 4b214948
This diff is collapsed.
doc/doc_ch/config.md
View file @ 4b214948
This diff is collapsed.
doc/doc_ch/models_list.md
View file @ 4b214948
This diff is collapsed.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment