SSD,Faster RCNN,YOLO V2

Based on my understanding, there are main 3 types of objection detection architecture as below:
- SSD
- Faster RCNN
- YOLO V2
 다음 글 감사합니다.
http://blog.csdn.net/u010167269/article/details/52563573
hard negative mining
data augmentation
ssd_mobilenet_v1_pets. config 설정 파일 은 다음 과 같 습 니 다.

# SSD with Mobilenet v1, configured for Oxford-IIIT Pets Dataset.
# Users should configure the fine_tune_checkpoint field in the train config as
# well as the label_map_path and input_path fields in the train_input_reader and
# eval_input_reader. Search for "PATH_TO_BE_CONFIGURED" to find the fields that
# should be configured.

model {
  ssd {
    num_classes: 1                 #      
    box_coder {                    
      faster_rcnn_box_coder {
        y_scale: 10.0
        x_scale: 10.0
        height_scale: 5.0
        width_scale: 5.0
      }
    }
    matcher {
      argmax_matcher {
        matched_threshold: 0.5
        unmatched_threshold: 0.5
        ignore_thresholds: false
        negatives_lower_than_unmatched: true
        force_match_for_each_row: true
      }
    }
    similarity_calculator {
      iou_similarity {
      }
    }
    anchor_generator {
      ssd_anchor_generator {
        num_layers: 6             # convolutional feature layers : 6 
        min_scale: 0.2            #             （aspect_ratios=1)    Sk=Smin+((Smax-Smin)/(m-1))*(k-1) 

        max_scale: 0.95           #  k:1...6; min_scale:Smin; max_scale:Smax 
        aspect_ratios: 1.0        #    =  =（0.2+（（0.95-0.2）/（6-1））*（k-1))*300   300       ；

                                 #    aspect_ratios=1；            ：Sk'=sqrt(Sk*Sk+1);

                                #       5 aspect_ratios;  6  

aspect_ratios: 2.0 \ # 계산 방법 은 동일 합 니 다. 상자 너비 만 = (0.2 + (0.95 - 0.2) / (6 - 1) * (k - 1) * sqrt (2) * 300 \ # 상자 높이 = (0.2 + (0.95 - 0.2) / (6 - 1) * (k - 1) / sqlt (2) * 300 aspect ratios: 0.5 aspect ratios: 3.0 aspect ratios: 0.3333} image resizer {fixed shape resizer {height: 300 width: 300} box predictor {convolutional box predictor{ min_depth: 0 max_depth: 0 num_layers_before_predictor: 0 use_dropout: false dropout_keep_probability: 0.8 kernel_size: 1 box_code_size: 4 apply_sigmoid_to_scores: false conv_hyperparams { activation: RELU_6, regularizer { l2_regularizer { weight: 0.00004 } } initializer { truncated_normal_initializer { stddev: 0.03 mean: 0.0 } } batch_norm { train: true, scale: true, center: true, decay: 0.9997, epsilon: 0.001, } } } } feature_extractor { type: 'ssd_mobilenet_v1' min_depth: 16 depth_multiplier: 1.0 conv_hyperparams { activation: RELU_6, regularizer { l2_regularizer { weight: 0.00004 } } initializer { truncated_normal_initializer { stddev: 0.03 mean: 0.0 } } batch_norm { train: true, scale: true, center: true, decay: 0.9997, epsilon: 0.001, } } } loss { classification_loss { weighted_sigmoid { anchorwise_output: true } } localization_loss { weighted_smooth_l1 { anchorwise_output: true } } hard_example_miner { num_hard_examples: 3000 iou_threshold: 0.99 loss_type: CLASSIFICATION max_negatives_per_positive: 3 min_negatives_per_image: 0 } classification_weight: 1.0 localization_weight: 1.0 } normalize_loss_by_num_matches: true post_processing { batch_non_max_suppression { score_threshold: 1e-8 iou_threshold: 0.6 max_detections_per_class: 100 max_total_detections: 100 } score_converter: SIGMOID } }}train_config: { batch_size: 5 optimizer { rms_prop_optimizer: { learning_rate: { exponential_decay_learning_rate { initial_learning_rate: 0.004 decay_steps: 800720 decay_factor: 0.95 } } momentum_optimizer_value: 0.9 decay: 0.9 epsilon: 1.0 } } fine_tune_checkpoint: "ssd_mobilenet_v1_coco_11_06_2017/model.ckpt" from_detection_checkpoint: true # Note: The below line limits the training process to 200K steps, which we # empirically found to be sufficient enough to train the pets dataset. This # effectively bypasses the learning rate schedule (the learning rate will # never decay). Remove the below line to train indefinitely. #num_steps: 200000 data_augmentation_options { random_horizontal_flip { } } data_augmentation_options { ssd_random_crop { } }}train_input_reader: { tf_record_input_reader { input_path: "data/train.record" } label_map_path: "data/object_detection.pbtxt"}eval_config: { num_examples: 40 # Note: The below line limits the evaluation process to 10 evaluations. # Remove the below line to evaluate indefinitely. #max_evals: 10}eval_input_reader: { tf_record_input_reader { input_path: "data/test.record" } label_map_path: "training/object_detection.pbtxt" shuffle: false num_readers: 1}
훈련 결 과 는 여러 가지 상황 이 발생 할 수 있다.