1、新增图片与视频活体检测

2、新增人脸属性识别（性别、年龄、口罩、姿态、眼睛状态）
3、优化检测返回与包结构
4、新增 dependencyManagement 统一依赖版本管理

smartjavaai-ocr/pom.xml

@@ -6,7 +6,7 @@
     <parent>
         <groupId>cn.smartjavaai</groupId>
         <artifactId>smartjavaai-parent</artifactId>
-        <version>1.0.11</version>
+        <version>1.0.12</version>
     </parent>
 
     <artifactId>smartjavaai-ocr</artifactId>
@@ -41,7 +41,7 @@
         </dependency>
     </dependencies>
 
-    <version>1.0.11</version>
+    <version>1.0.12</version>
     <name>smartjavaai-ocr</name>
     <description>SmartJavaAI</description>
     <url>https://github.com/geekwenjie/SmartJavaAI</url>

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/AbstractOcrModel.java

@@ -1,24 +0,0 @@
-package cn.smartjavaai.ocr;
-
-import cn.smartjavaai.common.entity.DetectionResponse;
-
-/**
- * 人脸识别算法
- * @author dwj
- */
-public abstract class AbstractOcrModel implements OcrModel {
-    @Override
-    public void loadModel(OcrModelConfig config) {
-        throw new UnsupportedOperationException("默认不支持该功能");
-    }
-
-    @Override
-    public DetectionResponse detect(String imagePath) {
-        throw new UnsupportedOperationException("默认不支持该功能");
-    }
-
-    @Override
-    public void detectAndDraw(String imagePath, String outputPath) {
-        throw new UnsupportedOperationException("默认不支持该功能");
-    }
-}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/detection/OcrDetModel.java

@@ -1,35 +1,37 @@
-package cn.smartjavaai.ocr;
+package cn.smartjavaai.ocr.detection;
 
 import cn.smartjavaai.common.entity.DetectionResponse;
 
-import java.awt.image.BufferedImage;
-import java.io.InputStream;
-
 /**
  * 人脸识别算法
  * @author dwj
  */
-public interface OcrModel {
+public interface OcrDetModel {
 
     /**
      * 加载模型
      * @param config
      */
-    void loadModel(OcrModelConfig config); // 加载模型
-
+    void loadModel(OcrDetModelConfig config); // 加载模型
 
     /**
      * 人脸检测
      * @param imagePath 图片路径
      * @return
      */
-    DetectionResponse detect(String imagePath);
+    default DetectionResponse detect(String imagePath) {
+        throw new UnsupportedOperationException("默认不支持该功能");
+    }
 
     /**
      * 检测并绘制结果
      * @param imagePath 图片输入路径（包含文件名称）
      * @param outputPath 图片输出路径（包含文件名称）
      */
-    void detectAndDraw(String imagePath, String outputPath);
+    default void detectAndDraw(String imagePath, String outputPath) {
+        throw new UnsupportedOperationException("默认不支持该功能");
+    }
+
+
 
 }

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/detection/OcrDetModelConfig.java

@@ -1,4 +1,4 @@
-package cn.smartjavaai.ocr;
+package cn.smartjavaai.ocr.detection;
 
 import cn.smartjavaai.common.enums.DeviceEnum;
 import lombok.Data;
@@ -8,12 +8,12 @@ import lombok.Data;
  * @date 2025/4/22
  */
 @Data
-public class OcrModelConfig {
+public class OcrDetModelConfig {
 
     /**
      * 模型名称
      */
-    private OcrModelEnum modelEnum;
+    private OcrDetModelEnum modelEnum;
 
     /**
      * 设备类型

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/detection/OcrDetModelEnum.java

@@ -1,11 +1,11 @@
-package cn.smartjavaai.ocr;
+package cn.smartjavaai.ocr.detection;
 
 /**
  * OCR模型枚举
  * @author dwj
  * @date 2025/4/4
  */
-public enum OcrModelEnum {
+public enum OcrDetModelEnum {
 
     PADDLEOCR_V4_DET_MODEL;
 
@@ -13,9 +13,9 @@ public enum OcrModelEnum {
     /**
      * 根据名称获取枚举 (忽略大小写和下划线变体)
      */
-    public static OcrModelEnum fromName(String name) {
+    public static OcrDetModelEnum fromName(String name) {
         String formatted = name.trim().toUpperCase().replaceAll("[-_]", "");
-        for (OcrModelEnum model : values()) {
+        for (OcrDetModelEnum model : values()) {
             if (model.name().replaceAll("_", "").equals(formatted)) {
                 return model;
             }

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/detection/OcrDetModelFactory.java

@@ -1,8 +1,8 @@
-package cn.smartjavaai.ocr;
+package cn.smartjavaai.ocr.detection;
 
 import cn.smartjavaai.common.config.Config;
 import cn.smartjavaai.ocr.exception.OcrException;
-import cn.smartjavaai.ocr.model.PaddleOCRV4DetectModel;
+import cn.smartjavaai.ocr.ppv4.model.PaddleOCRV4DetModel;
 import lombok.extern.slf4j.Slf4j;
 
 import java.util.Map;
@@ -14,25 +14,25 @@ import java.util.concurrent.ConcurrentHashMap;
  * @author dwj
  */
 @Slf4j
-public class OcrModelFactory {
+public class OcrDetModelFactory {
 
     // 使用 volatile 和双重检查锁定来确保线程安全的单例模式
-    private static volatile OcrModelFactory instance;
+    private static volatile OcrDetModelFactory instance;
 
-    private static final ConcurrentHashMap<String, OcrModel> modelMap = new ConcurrentHashMap<>();
+    private static final ConcurrentHashMap<String, OcrDetModel> modelMap = new ConcurrentHashMap<>();
 
     /**
      * 算法注册表
      */
-    private static final Map<String, Class<? extends OcrModel>> registry =
+    private static final Map<String, Class<? extends OcrDetModel>> registry =
             new ConcurrentHashMap<>();
 
 
-    public static OcrModelFactory getInstance() {
+    public static OcrDetModelFactory getInstance() {
         if (instance == null) {
-            synchronized (OcrModelFactory.class) {
+            synchronized (OcrDetModelFactory.class) {
                 if (instance == null) {
-                    instance = new OcrModelFactory();
+                    instance = new OcrDetModelFactory();
                 }
             }
         }
@@ -46,7 +46,7 @@ public class OcrModelFactory {
      * @param name
      * @param clazz
      */
-    private static void registerModel(String name, Class<? extends OcrModel> clazz) {
+    private static void registerModel(String name, Class<? extends OcrDetModel> clazz) {
         registry.put(name.toLowerCase(), clazz);
     }
 
@@ -56,7 +56,7 @@ public class OcrModelFactory {
      * @param config
      * @return
      */
-    public OcrModel getModel(OcrModelConfig config) {
+    public OcrDetModel getModel(OcrDetModelConfig config) {
         if(Objects.isNull(config) || Objects.isNull(config.getModelEnum())){
             throw new OcrException("未配置OCR模型");
         }
@@ -70,14 +70,14 @@ public class OcrModelFactory {
      * @param config
      * @return
      */
-    private OcrModel createFaceModel(OcrModelConfig config) {
+    private OcrDetModel createFaceModel(OcrDetModelConfig config) {
         Class<?> clazz = registry.get(config.getModelEnum().name().toLowerCase());
         if(clazz == null){
             throw new OcrException("Unsupported model");
         }
-        OcrModel algorithm = null;
+        OcrDetModel algorithm = null;
         try {
-            algorithm = (OcrModel) clazz.newInstance();
+            algorithm = (OcrDetModel) clazz.newInstance();
         } catch (InstantiationException | IllegalAccessException e) {
             throw new OcrException(e);
         }
@@ -88,7 +88,7 @@ public class OcrModelFactory {
 
     // 初始化默认算法
     static {
-        registerModel("PADDLEOCR_V4_DET_MODEL", PaddleOCRV4DetectModel.class);
+        registerModel("PADDLEOCR_V4_DET_MODEL", PaddleOCRV4DetModel.class);
         log.info("缓存目录：{}", Config.getCachePath());
     }
 

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/model/PaddleOCRV4Model.java

@@ -1,27 +0,0 @@
-package cn.smartjavaai.ocr.model;
-
-import ai.djl.inference.Predictor;
-import ai.djl.modality.cv.Image;
-import ai.djl.ndarray.NDList;
-import ai.djl.ndarray.NDManager;
-import ai.djl.repository.zoo.Criteria;
-import ai.djl.repository.zoo.ModelZoo;
-import ai.djl.repository.zoo.ZooModel;
-import ai.djl.training.util.ProgressBar;
-import cn.smartjavaai.ocr.translator.PaddleOCRV4DetectionTranslator;
-import org.opencv.core.Mat;
-
-import java.nio.file.Paths;
-import java.util.concurrent.ConcurrentHashMap;
-
-/**
- * @author dwj
- * @date 2025/4/21
- */
-public class PaddleOCRV4Model {
-
-
-    public void loadModel(){
-
-    }
-}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/ppv4/model/PaddleOCRV4DetModel.java

@@ -1,25 +1,22 @@
-package cn.smartjavaai.ocr.model;
+package cn.smartjavaai.ocr.ppv4.model;
 
 import ai.djl.MalformedModelException;
 import ai.djl.inference.Predictor;
 import ai.djl.modality.cv.Image;
 import ai.djl.modality.cv.ImageFactory;
-import ai.djl.modality.cv.output.DetectedObjects;
 import ai.djl.ndarray.NDList;
-import ai.djl.ndarray.NDManager;
 import ai.djl.repository.zoo.Criteria;
 import ai.djl.repository.zoo.ModelNotFoundException;
 import ai.djl.repository.zoo.ModelZoo;
 import ai.djl.repository.zoo.ZooModel;
 import ai.djl.training.util.ProgressBar;
-import ai.djl.translate.TranslateException;
 import cn.smartjavaai.common.entity.DetectionResponse;
 import cn.smartjavaai.common.pool.PredictorFactory;
 import cn.smartjavaai.common.utils.FileUtils;
-import cn.smartjavaai.ocr.AbstractOcrModel;
-import cn.smartjavaai.ocr.OcrModelConfig;
+import cn.smartjavaai.ocr.detection.OcrDetModelConfig;
 import cn.smartjavaai.ocr.exception.OcrException;
-import cn.smartjavaai.ocr.translator.PaddleOCRV4DetectionTranslator;
+import cn.smartjavaai.ocr.detection.OcrDetModel;
+import cn.smartjavaai.ocr.ppv4.translator.PaddleOCRV4DetectTranslator;
 import cn.smartjavaai.ocr.utils.ImageUtils;
 import cn.smartjavaai.ocr.utils.OcrUtils;
 import lombok.extern.slf4j.Slf4j;
@@ -40,14 +37,14 @@ import java.util.concurrent.ConcurrentHashMap;
  * @date 2025/4/21
  */
 @Slf4j
-public class PaddleOCRV4DetectModel extends AbstractOcrModel {
+public class PaddleOCRV4DetModel implements OcrDetModel {
 
     private ZooModel detectionModel;
 
     private ObjectPool<Predictor<Image, NDList>> predictorPool;
 
     @Override
-    public void loadModel(OcrModelConfig config){
+    public void loadModel(OcrDetModelConfig config){
         if(StringUtils.isBlank(config.getModelPath())){
             throw new OcrException("modelPath is null");
         }
@@ -56,7 +53,7 @@ public class PaddleOCRV4DetectModel extends AbstractOcrModel {
                         .optEngine("OnnxRuntime")
                         .setTypes(Image.class, NDList.class)
                         .optModelPath(Paths.get(config.getModelPath()))
-                        .optTranslator(new PaddleOCRV4DetectionTranslator(new ConcurrentHashMap<String, String>()))
+                        .optTranslator(new PaddleOCRV4DetectTranslator(new ConcurrentHashMap<String, String>()))
                         .optProgress(new ProgressBar())
                         .build();
         try{
@@ -83,11 +80,6 @@ public class PaddleOCRV4DetectModel extends AbstractOcrModel {
         return detect(img);
     }
 
-    /**
-     * 人脸检测
-     * @param image
-     * @return
-     */
     private DetectionResponse detect(Image image){
         Predictor<Image, NDList> predictor = null;
         try {
@@ -120,7 +112,7 @@ public class PaddleOCRV4DetectModel extends AbstractOcrModel {
         try {
             Image img = ImageFactory.getInstance().fromFile(Paths.get(imagePath));
             DetectionResponse detectionResponse = detect(img);
-            if(Objects.isNull(detectionResponse) || Objects.isNull(detectionResponse.getRectangleList()) || detectionResponse.getRectangleList().isEmpty()){
+            if(Objects.isNull(detectionResponse) || Objects.isNull(detectionResponse.getDetectionInfoList()) || detectionResponse.getDetectionInfoList().isEmpty()){
                 throw new OcrException("未识别到文字");
             }
             ImageUtils.drawRect((Mat)img.getWrappedImage(), detectionResponse);

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/ppv4/model/PaddleOCRV4RecModel.java

@@ -0,0 +1,129 @@
+package cn.smartjavaai.ocr.ppv4.model;
+
+import ai.djl.MalformedModelException;
+import ai.djl.inference.Predictor;
+import ai.djl.modality.cv.Image;
+import ai.djl.modality.cv.ImageFactory;
+import ai.djl.ndarray.NDList;
+import ai.djl.repository.zoo.Criteria;
+import ai.djl.repository.zoo.ModelNotFoundException;
+import ai.djl.repository.zoo.ModelZoo;
+import ai.djl.repository.zoo.ZooModel;
+import ai.djl.training.util.ProgressBar;
+import cn.smartjavaai.common.entity.DetectionResponse;
+import cn.smartjavaai.common.pool.PredictorFactory;
+import cn.smartjavaai.common.utils.FileUtils;
+import cn.smartjavaai.ocr.detection.OcrDetModel;
+import cn.smartjavaai.ocr.detection.OcrDetModelConfig;
+import cn.smartjavaai.ocr.exception.OcrException;
+import cn.smartjavaai.ocr.ppv4.translator.PaddleOCRV4DetectTranslator;
+import cn.smartjavaai.ocr.recognition.OcrRecModel;
+import cn.smartjavaai.ocr.recognition.OcrRecModelConfig;
+import cn.smartjavaai.ocr.utils.ImageUtils;
+import cn.smartjavaai.ocr.utils.OcrUtils;
+import lombok.extern.slf4j.Slf4j;
+import org.apache.commons.lang3.StringUtils;
+import org.apache.commons.pool2.ObjectPool;
+import org.apache.commons.pool2.impl.GenericObjectPool;
+import org.opencv.core.Mat;
+
+import java.io.IOException;
+import java.nio.file.Files;
+import java.nio.file.Path;
+import java.nio.file.Paths;
+import java.util.Objects;
+import java.util.concurrent.ConcurrentHashMap;
+
+/**
+ * PaddleOCRV4 识别模型实现
+ * @author dwj
+ * @date 2025/4/21
+ */
+@Slf4j
+public class PaddleOCRV4RecModel implements OcrRecModel {
+
+    private ZooModel detectionModel;
+
+    private ObjectPool<Predictor<Image, NDList>> predictorPool;
+
+    @Override
+    public void loadModel(OcrRecModelConfig config){
+        if(StringUtils.isBlank(config.getModelPath())){
+            throw new OcrException("modelPath is null");
+        }
+        Criteria<Image, NDList> criteria =
+                Criteria.builder()
+                        .optEngine("OnnxRuntime")
+                        .setTypes(Image.class, NDList.class)
+                        .optModelPath(Paths.get(config.getModelPath()))
+                        .optTranslator(new PaddleOCRV4DetectTranslator(new ConcurrentHashMap<String, String>()))
+                        .optProgress(new ProgressBar())
+                        .build();
+        try{
+            detectionModel = ModelZoo.loadModel(criteria);
+            // 创建池子：每个线程独享 Predictor
+            this.predictorPool = new GenericObjectPool<>(new PredictorFactory<>(detectionModel));
+            log.info("当前设备: " + detectionModel.getNDManager().getDevice());
+        } catch (IOException | ModelNotFoundException | MalformedModelException e) {
+            throw new OcrException("模型加载失败", e);
+        }
+    }
+
+    @Override
+    public DetectionResponse detect(String imagePath){
+        if(!FileUtils.isFileExists(imagePath)){
+            throw new OcrException("图像文件不存在");
+        }
+        Image img = null;
+        try {
+            img = ImageFactory.getInstance().fromFile(Paths.get(imagePath));
+        } catch (IOException e) {
+            throw new OcrException("无效的图片", e);
+        }
+        return detect(img);
+    }
+
+    private DetectionResponse detect(Image image){
+        Predictor<Image, NDList> predictor = null;
+        try {
+            predictor = predictorPool.borrowObject();
+            NDList result = predictor.predict(image);
+            return OcrUtils.convertToDetectionResponse(result, image);
+        } catch (Exception e) {
+            throw new OcrException("OCR检测错误", e);
+        }finally {
+            if (predictor != null) {
+                try {
+                    predictorPool.returnObject(predictor); //归还
+                } catch (Exception e) {
+                    log.warn("归还Predictor失败", e);
+                    try {
+                        predictor.close(); // 归还失败才销毁
+                    } catch (Exception ex) {
+                        log.error("关闭Predictor失败", ex);
+                    }
+                }
+            }
+        }
+    }
+
+    @Override
+    public void detectAndDraw(String imagePath, String outputPath) {
+        if(!FileUtils.isFileExists(imagePath)){
+            throw new OcrException("图像文件不存在");
+        }
+        try {
+            Image img = ImageFactory.getInstance().fromFile(Paths.get(imagePath));
+            DetectionResponse detectionResponse = detect(img);
+            if(Objects.isNull(detectionResponse) || Objects.isNull(detectionResponse.getDetectionInfoList()) || detectionResponse.getDetectionInfoList().isEmpty()){
+                throw new OcrException("未识别到文字");
+            }
+            ImageUtils.drawRect((Mat)img.getWrappedImage(), detectionResponse);
+            Path output = Paths.get(outputPath);
+            log.info("Saving to {}", output.toAbsolutePath().toString());
+            img.save(Files.newOutputStream(output), "png");
+        } catch (IOException e) {
+            throw new OcrException(e);
+        }
+    }
+}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/ppv4/translator/PaddleOCRV4DetectTranslator.java

@@ -1,4 +1,4 @@
-package cn.smartjavaai.ocr.translator;
+package cn.smartjavaai.ocr.ppv4.translator;
 
 import ai.djl.modality.cv.Image;
 import ai.djl.modality.cv.util.NDImageUtils;
@@ -27,7 +27,7 @@ import java.util.Map;
  * @mail 179209347@qq.com
  * @website www.aias.top
  */
-public class PaddleOCRV4DetectionTranslator implements Translator<Image, NDList> {
+public class PaddleOCRV4DetectTranslator implements Translator<Image, NDList> {
     // det_algorithm == "DB"
     private final float thresh = 0.3f;
     private final boolean use_dilation = false;
@@ -44,7 +44,7 @@ public class PaddleOCRV4DetectionTranslator implements Translator<Image, NDList>
     private int img_height;
     private int img_width;
 
-    public PaddleOCRV4DetectionTranslator(Map<String, ?> arguments) {
+    public PaddleOCRV4DetectTranslator(Map<String, ?> arguments) {
         limit_side_len =
                 arguments.containsKey("limit_side_len")
                         ? Integer.parseInt(arguments.get("limit_side_len").toString())

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/ppv4/translator/PaddleOCRV4WordRecTranslator.java

@@ -1,4 +1,4 @@
-package cn.smartjavaai.ocr.translator;
+package cn.smartjavaai.ocr.ppv4.translator;
 
 import ai.djl.Model;
 import ai.djl.modality.cv.Image;

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/recognition/OcrRecModel.java

@@ -0,0 +1,37 @@
+package cn.smartjavaai.ocr.recognition;
+
+import cn.smartjavaai.common.entity.DetectionResponse;
+
+/**
+ * OCR模型
+ * @author dwj
+ */
+public interface OcrRecModel {
+
+    /**
+     * 加载模型
+     * @param config
+     */
+    void loadModel(OcrRecModelConfig config); // 加载模型
+
+    /**
+     * 人脸检测
+     * @param imagePath 图片路径
+     * @return
+     */
+    default DetectionResponse detect(String imagePath) {
+        throw new UnsupportedOperationException("默认不支持该功能");
+    }
+
+    /**
+     * 检测并绘制结果
+     * @param imagePath 图片输入路径（包含文件名称）
+     * @param outputPath 图片输出路径（包含文件名称）
+     */
+    default void detectAndDraw(String imagePath, String outputPath) {
+        throw new UnsupportedOperationException("默认不支持该功能");
+    }
+
+
+
+}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/recognition/OcrRecModelConfig.java

@@ -0,0 +1,28 @@
+package cn.smartjavaai.ocr.recognition;
+
+import cn.smartjavaai.common.enums.DeviceEnum;
+import lombok.Data;
+
+/**
+ * @author dwj
+ * @date 2025/4/22
+ */
+@Data
+public class OcrRecModelConfig {
+
+    /**
+     * 模型名称
+     */
+    private OcrRecModelEnum modelEnum;
+
+    /**
+     * 设备类型
+     */
+    private DeviceEnum device;
+
+    /**
+     * 模型路径
+     */
+    private String modelPath;
+
+}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/recognition/OcrRecModelEnum.java

@@ -0,0 +1,27 @@
+package cn.smartjavaai.ocr.recognition;
+
+/**
+ * OCR识别模型枚举
+ * @author dwj
+ * @date 2025/4/4
+ */
+public enum OcrRecModelEnum {
+
+    PADDLEOCR_V4_REC_MODEL;
+
+
+    /**
+     * 根据名称获取枚举 (忽略大小写和下划线变体)
+     */
+    public static OcrRecModelEnum fromName(String name) {
+        String formatted = name.trim().toUpperCase().replaceAll("[-_]", "");
+        for (OcrRecModelEnum model : values()) {
+            if (model.name().replaceAll("_", "").equals(formatted)) {
+                return model;
+            }
+        }
+        throw new IllegalArgumentException("未知模型名称: " + name);
+    }
+
+
+}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/recognition/OcrRecModelFactory.java

@@ -0,0 +1,95 @@
+package cn.smartjavaai.ocr.recognition;
+
+import cn.smartjavaai.common.config.Config;
+import cn.smartjavaai.ocr.exception.OcrException;
+import cn.smartjavaai.ocr.ppv4.model.PaddleOCRV4DetModel;
+import lombok.extern.slf4j.Slf4j;
+
+import java.util.Map;
+import java.util.Objects;
+import java.util.concurrent.ConcurrentHashMap;
+
+/**
+ * OCR模型工厂
+ * @author dwj
+ */
+@Slf4j
+public class OcrRecModelFactory {
+
+    // 使用 volatile 和双重检查锁定来确保线程安全的单例模式
+    private static volatile OcrRecModelFactory instance;
+
+    private static final ConcurrentHashMap<String, OcrRecModel> modelMap = new ConcurrentHashMap<>();
+
+    /**
+     * 算法注册表
+     */
+    private static final Map<String, Class<? extends OcrRecModel>> registry =
+            new ConcurrentHashMap<>();
+
+
+    public static OcrRecModelFactory getInstance() {
+        if (instance == null) {
+            synchronized (OcrRecModelFactory.class) {
+                if (instance == null) {
+                    instance = new OcrRecModelFactory();
+                }
+            }
+        }
+        return instance;
+    }
+
+
+
+    /**
+     * 注册算法
+     * @param name
+     * @param clazz
+     */
+    private static void registerModel(String name, Class<? extends OcrRecModel> clazz) {
+        registry.put(name.toLowerCase(), clazz);
+    }
+
+
+    /**
+     * 获取模型（通过配置）
+     * @param config
+     * @return
+     */
+    public OcrRecModel getModel(OcrRecModelConfig config) {
+        if(Objects.isNull(config) || Objects.isNull(config.getModelEnum())){
+            throw new OcrException("未配置OCR模型");
+        }
+        return modelMap.computeIfAbsent(config.getModelEnum().name(), k -> {
+            return createFaceModel(config);
+        });
+    }
+
+    /**
+     * 使用ModelConfig创建算法
+     * @param config
+     * @return
+     */
+    private OcrRecModel createFaceModel(OcrRecModelConfig config) {
+        Class<?> clazz = registry.get(config.getModelEnum().name().toLowerCase());
+        if(clazz == null){
+            throw new OcrException("Unsupported model");
+        }
+        OcrRecModel algorithm = null;
+        try {
+            algorithm = (OcrRecModel) clazz.newInstance();
+        } catch (InstantiationException | IllegalAccessException e) {
+            throw new OcrException(e);
+        }
+        algorithm.loadModel(config);
+        return algorithm;
+    }
+
+
+    // 初始化默认算法
+    static {
+        //registerModel("PADDLEOCR_V4_REC_MODEL", PaddleOCRV4DetModel.class);
+        log.info("缓存目录：{}", Config.getCachePath());
+    }
+
+}

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/utils/ImageUtils.java

@@ -4,6 +4,7 @@ import ai.djl.modality.cv.Image;
 import ai.djl.modality.cv.ImageFactory;
 import ai.djl.modality.cv.output.DetectedObjects;
 import ai.djl.ndarray.NDArray;
+import cn.smartjavaai.common.entity.DetectionInfo;
 import cn.smartjavaai.common.entity.DetectionRectangle;
 import cn.smartjavaai.common.entity.DetectionResponse;
 import cn.smartjavaai.common.utils.OpenCVUtils;
@@ -207,7 +208,8 @@ public class ImageUtils {
      */
     public static void drawRect(Mat mat, DetectionResponse detectionResponse) {
 
-        for(DetectionRectangle detectionRectangle : detectionResponse.getRectangleList()){
+        for(DetectionInfo detectionInfo : detectionResponse.getDetectionInfoList()){
+            DetectionRectangle detectionRectangle = detectionInfo.getDetectionRectangle();
             // 左上角点
             Point topLeft = new Point(detectionRectangle.getX(), detectionRectangle.getY());
             // 右下角点

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/utils/OcrUtils.java

@@ -5,6 +5,7 @@ import ai.djl.modality.cv.output.BoundingBox;
 import ai.djl.modality.cv.output.DetectedObjects;
 import ai.djl.ndarray.NDArray;
 import ai.djl.ndarray.NDList;
+import cn.smartjavaai.common.entity.DetectionInfo;
 import cn.smartjavaai.common.entity.DetectionRectangle;
 import cn.smartjavaai.common.entity.DetectionResponse;
 import lombok.extern.slf4j.Slf4j;
@@ -34,7 +35,7 @@ public class OcrUtils {
             return null;
         }
         DetectionResponse detectionResponse = new DetectionResponse();
-        List<DetectionRectangle> rectangleList = new ArrayList<DetectionRectangle>();
+        List<DetectionInfo> detectionInfoList = new ArrayList<DetectionInfo>();
         for(NDArray box : dt_boxes){
             DetectionRectangle rectangle = new DetectionRectangle();
             float[] points = box.toFloatArray();
@@ -54,9 +55,9 @@ public class OcrUtils {
             rectangle.setY(y);
             rectangle.setHeight(height);
             rectangle.setWidth(width);
-            rectangleList.add(rectangle);
+            detectionInfoList.add(new DetectionInfo(rectangle));
         }
-        detectionResponse.setRectangleList(rectangleList);
+        detectionResponse.setDetectionInfoList(detectionInfoList);
         return detectionResponse;
     }