核心流程：

1. 收集人脸图片
2. 处理图片，并按脸型打标（长型脸、圆形脸、椭圆脸、方形脸、心形脸）
3. PC端使用Python + Tensorflow进行脸型模型训练
4. 转换为Android可用模型
5. 手机端使用OpenCV + Tensorflow Lite进行人脸检测和脸型分析

Tensorflow模型训练核心代码：

model = tf.keras.models.Sequential([
    # Note the input shape is the desired size of the image 190x250 with 1 bytes color
    # This is the first convolution
    tf.keras.layers.Conv2D(64, (3, 3), activation='relu', input_shape=(250, 190, 1)),
    tf.keras.layers.MaxPooling2D(2, 2),
    # The second convolution
    tf.keras.layers.Conv2D(64, (3, 3), activation='relu'),
    tf.keras.layers.MaxPooling2D(2, 2),
    # The third convolution
    tf.keras.layers.Conv2D(128, (3, 3), activation='relu'),
    tf.keras.layers.MaxPooling2D(2, 2),
    # The fourth convolution
    tf.keras.layers.Conv2D(128, (3, 3), activation='relu'),
    tf.keras.layers.MaxPooling2D(2, 2),
    # Flatten the results to feed into a DNN
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dropout(0.3),
    # 512 neuron hidden layer
    tf.keras.layers.Dense(512, activation='relu'),
    tf.keras.layers.Dense(5, activation='softmax')
])

# Print the model summary
model.summary()

# Set the training parameters
model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])

TRAINING_DIR = '''../dataset/train'''
training_datagen = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1. / 255)

VALIDATION_DIR = '''../dataset/test'''
validation_datagen = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1. / 255)

Android端核心代码：

public int predictImage(Bitmap bitmap) {
   // 将图像转换为TensorFlow Lite期望的格式
    ByteBuffer inputBuffer = preprocessImage(bitmap);

    // 进行推理
    float[][] output = new float[1][NUM_CLASSES];
    interpreter.run(inputBuffer, output);

    // 获取预测类别
    return argmax(output[0]);
}

// 初始化脸型检测模型
imageClassifier = new ImageClassifier(FaceShapeWithTensorflowActivity.this, "model/model01.tflite", MODEL_WIDTH, MODEL_HEIGHT, 1, 5);

//开始识别
if(lastFaceNo == -1 && imageClassifier != null)
    lastFaceNo = imageClassifier.predictImage(bitmap);
int thisFaceNo = imageClassifier.predictImage(bitmap);
if(imageClassifier != null && lastFaceNo == thisFaceNo) { //检测同一个脸型
    failGetCnt = 0;
    successGetCnt++;
    if(successGetCnt > FaceConstant.FACE_INFO_RECOGNIZE_CNT) {
        runOnUiThread(new Runnable() {
            @Override
            public void run() {
                //显示裁切出来的人脸图片
                imgGetFace.setImageBitmap(bitmap);
                imgFaceTipsBg.setVisibility(View.VISIBLE);
                imgFaceTipsBg.setImageResource(R.mipmap.ic_tips_gate_success);
                tvFaceTips.setText(faceShapeMap.get(thisFaceNo));
                speechText(faceShapeMap.get(thisFaceNo));
            }
        });
    }
} else {
    failGetCnt++;
    if(failGetCnt > FaceConstant.FACE_INFO_ALLOW_RECOGNIZE_FAIL_CNT) {
        successGetCnt = 0;
        runOnUiThread(new Runnable() {
            @Override
            public void run() {
                imgFaceTipsBg.setVisibility(View.VISIBLE);
                imgFaceTipsBg.setImageResource(R.mipmap.ic_tips_gate_fail);
                tvFaceTips.setText("识别失败");
            }
        });
    }
}
lastFaceNo = thisFaceNo;

有需要源码的同学可以私信我！！！！！！！

如果能帮到您，您又刚好方便，又刚好打开了微信，那就鼓励下咔嚓小豪哥我吖