mv_machine_learning/face_recognition/include/face_recognition.h

   1 /**
   2  * Copyright (c) 2022 Samsung Electronics Co., Ltd All Rights Reserved
   3  *
   4  * Licensed under the Apache License, Version 2.0 (the "License");
   5  * you may not use this file except in compliance with the License.
   6  * You may obtain a copy of the License at
   7  *
   8  * http://www.apache.org/licenses/LICENSE-2.0
   9  *
  10  * Unless required by applicable law or agreed to in writing, software
  11  * distributed under the License is distributed on an "AS IS" BASIS,
  12  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13  * See the License for the specific language governing permissions and
  14  * limitations under the License.
  15  */
  16
  17 #ifndef __FACE_RECOGNITION_H__
  18 #define __FACE_RECOGNITION_H__
  19
  20 #include <unordered_map>
  21 #include <mv_common.h>
  22 #include <mv_inference_type.h>
  23
  24 #include "training_engine_error.h"
  25 #include "training_engine_common_impl.h"
  26 #include "inference_engine_common_impl.h"
  27 #include "Inference.h"
  28 #include "label_manager.h"
  29 #include "backbone_model_info.h"
  30 #include "feature_vector_manager.h"
  31 #include "data_set_manager.h"
  32 #include "simple_shot.h"
  33
  34 namespace mediavision
  35 {
  36 namespace machine_learning
  37 {
  38 namespace face_recognition
  39 {
  40 namespace status
  41 {
  42 enum { NONE = 0, INITIALIZED, REGISTERED, INFERENCED, DELETED };
  43 } // status
  44
  45 namespace mode
  46 {
  47 enum { REGISTER = 0, INFERENCE, DELETE };
  48 } // mode
  49
  50 } // face_recognition
  51
  52 struct face_recognition_register_input_s {
  53         std::vector<float> data;
  54         std::string label;
  55 };
  56
  57 struct mv_face_recognition_input_s {
  58         unsigned int mode;
  59         std::vector<face_recognition_register_input_s> register_src;
  60         std::vector<std::vector<float> > inputs;
  61         std::vector<std::string> labels;
  62 };
  63
  64 /**
  65  * @brief The face recognition result structure.
  66  * @details Contains face recognition result such as label, label index, raw data,
  67  *          and raw data count.
  68  */
  69 struct mv_face_recognition_result_s {
  70         unsigned int label_idx; /**< label index of label file. */
  71         std::vector<float> raw_data; /**< raw data to each label. */
  72         std::string label; /**< label string. */
  73         bool is_valid; /**< inference result is valid or not. */
  74 };
  75
  76 struct FaceRecognitionConfig {
  77         training_target_type_e training_target_device_type;
  78         training_backend_type_e training_engine_backend_type;
  79         mv_inference_target_device_e inference_target_device_type;
  80         mv_inference_backend_type_e inference_engine_backend_type;
  81         std::string internal_model_file_path;
  82         std::string label_file_path;
  83         std::string feature_vector_file_path;
  84         double decision_threshold;
  85         std::vector<size_t> input_tensor_shape;
  86 };
  87
  88 class FaceRecognition
  89 {
  90 private:
  91         unsigned int _status;
  92         std::unique_ptr<mediavision::inference::Inference> _internal;
  93         std::unique_ptr<mediavision::inference::Inference> _backbone;
  94         std::unique_ptr<IBackboneModelInfo> _backbone_model_info;
  95         std::unique_ptr<TrainingModel> _training_model;
  96         std::unique_ptr<LabelManager> _label_manager;
  97         FaceRecognitionConfig _config;
  98         mv_face_recognition_result_s _result;
  99
 100         // FYI. This function should be called every time a new face is registered.
 101         void ImportLabel();
 102         void CheckFeatureVectorFile(std::string fv_file_name, std::string new_fv_file_name);
 103         void StoreDataSet(std::unique_ptr<DataSetManager> &data_set, unsigned int label_cnt);
 104         int GetAnswer();
 105         std::vector<model_layer_info> &GetBackboneInputLayerInfo();
 106         int GetVecFromMvSource(mv_source_h img_src, std::vector<float> &out_vec);
 107
 108 public:
 109         FaceRecognition();
 110         ~FaceRecognition();
 111
 112         int Initialize();
 113         void SetConfig(FaceRecognitionConfig &config);
 114         int RegisterNewFace(std::vector<float> &input_vec, std::string label_name);
 115         int RecognizeFace(std::vector<float> &input_vec);
 116         int DeleteLabel(std::string label_name);
 117         int GetLabel(const char **out_label);
 118         mv_face_recognition_result_s &result();
 119 };
 120
 121 } // machine_learning
 122 } // mediavision
 123
 124 #endif