model exporter + raccoon detection demo

Author: abdelrahman-gaber

README.md

@@ -13,7 +13,7 @@ def DetectFromVideo(detector, Video_path, save_output=False, output_dir='output/
 		output_path = os.path.join(output_dir, 'detection_'+ Video_path.split("/")[-1])
 		frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
 		frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
-		out = cv2.VideoWriter(output_path, cv2.VideoWriter_fourcc(*"mp4v"), 25, (frame_width, frame_height))
+		out = cv2.VideoWriter(output_path, cv2.VideoWriter_fourcc(*"mp4v"), 30, (frame_width, frame_height))
 
 	while (cap.isOpened()):
 		ret, img = cap.read()
@@ -65,7 +65,7 @@ def DetectImagesFromFolder(detector, images_dir, save_output=False, output_dir='
 	parser.add_argument('--threshold', help='Detection Threshold', type=float, default=0.4)
 	parser.add_argument('--images_dir', help='Directory to input images)', default='data/samples/images/')
 	parser.add_argument('--video_path', help='Path to input video)', default='data/samples/pedestrian_test.mp4')
-	parser.add_argument('--output_directory', help='Path to output images and video', default='data/output')
+	parser.add_argument('--output_directory', help='Path to output images and video', default='data/samples/output')
 	parser.add_argument('--video_input', help='Flag for video input, default: False', action='store_true')  # default is false
 	parser.add_argument('--save_output', help='Flag for save images and video with detections visualized, default: False',
 	                    action='store_true')  # default is false

data/samples/clean_raccoon.mp4

@@ -58,13 +58,14 @@ def DisplayDetections(self, image, boxes_list, det_time=None):
 			score = str(np.round(boxes_list[idx][-1], 2))
 
 			text = cls + ": " + score
-			cv2.rectangle(img, (x_min, y_min), (x_max, y_max), (0, 255, 0), 2)
+			cv2.rectangle(img, (x_min, y_min), (x_max, y_max), (0, 255, 0), 1)
 			cv2.rectangle(img, (x_min, y_min - 20), (x_min, y_min), (255, 255, 255), -1)
-			cv2.putText(img, text, (x_min + 5, y_min - 7), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 0), 1)
+			cv2.putText(img, text, (x_min + 5, y_min - 7), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 1)
 
 		if det_time != None:
-			elapsed_time = str(det_time) + " ms"
-			cv2.putText(img, elapsed_time, (25, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 0, 0), 2)
+			fps = round(1000. / det_time, 1)
+			fps_txt = str(fps) + " FPS"
+			cv2.putText(img, fps_txt, (25, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 0, 0), 2)
 
 		return img
 

data/samples/docs/detection-clean-raccoon.gif

@@ -0,0 +1,4 @@
+model_checkpoint_path: "ckpt-0"
+all_model_checkpoint_paths: "ckpt-0"
+all_model_checkpoint_timestamps: 1595859041.5022295
+last_preserved_timestamp: 1595859040.4026093

detect_objects.py

@@ -0,0 +1,186 @@
+model {
+  ssd {
+    num_classes: 1
+    image_resizer {
+      fixed_shape_resizer {
+        height: 300
+        width: 300
+      }
+    }
+    feature_extractor {
+      type: "ssd_mobilenet_v2_keras"
+      depth_multiplier: 1.0
+      min_depth: 16
+      conv_hyperparams {
+        regularizer {
+          l2_regularizer {
+            weight: 4e-05
+          }
+        }
+        initializer {
+          truncated_normal_initializer {
+            mean: 0.0
+            stddev: 0.03
+          }
+        }
+        activation: RELU_6
+        batch_norm {
+          decay: 0.97
+          center: true
+          scale: true
+          epsilon: 0.001
+          train: true
+        }
+      }
+      override_base_feature_extractor_hyperparams: true
+    }
+    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: 0.4
+        unmatched_threshold: 0.4
+        ignore_thresholds: false
+        negatives_lower_than_unmatched: true
+        force_match_for_each_row: true
+        use_matmul_gather: true
+      }
+    }
+    similarity_calculator {
+      iou_similarity {
+      }
+    }
+    box_predictor {
+      convolutional_box_predictor {
+        conv_hyperparams {
+          regularizer {
+            l2_regularizer {
+              weight: 4e-05
+            }
+          }
+          initializer {
+            random_normal_initializer {
+              mean: 0.0
+              stddev: 0.01
+            }
+          }
+          activation: RELU_6
+          batch_norm {
+            decay: 0.97
+            center: true
+            scale: true
+            epsilon: 0.001
+            train: true
+          }
+        }
+        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
+        class_prediction_bias_init: -4.6
+      }
+    }
+    anchor_generator {
+      ssd_anchor_generator {
+        num_layers: 6
+        min_scale: 0.15
+        max_scale: 0.95
+        aspect_ratios: 1.0
+        aspect_ratios: 2.0
+        aspect_ratios: 0.5
+      }
+    }
+    post_processing {
+      batch_non_max_suppression {
+        score_threshold: 0.001
+        iou_threshold: 0.4
+        max_detections_per_class: 100
+        max_total_detections: 100
+      }
+      score_converter: SIGMOID
+    }
+    normalize_loss_by_num_matches: true
+    loss {
+      localization_loss {
+        weighted_smooth_l1 {
+          delta: 1.0
+        }
+      }
+      classification_loss {
+        weighted_sigmoid_focal {
+          gamma: 2.0
+          alpha: 0.75
+        }
+      }
+      classification_weight: 1.0
+      localization_weight: 1.0
+    }
+    encode_background_as_zeros: true
+    normalize_loc_loss_by_codesize: true
+    inplace_batchnorm_update: true
+    freeze_batchnorm: false
+  }
+}
+train_config {
+  batch_size: 16
+  data_augmentation_options {
+    random_horizontal_flip {
+    }
+  }
+  data_augmentation_options {
+    ssd_random_crop {
+    }
+  }
+  sync_replicas: true
+  optimizer {
+    momentum_optimizer {
+      learning_rate {
+        cosine_decay_learning_rate {
+          learning_rate_base: 0.025
+          total_steps: 3000
+          warmup_learning_rate: 0.005
+          warmup_steps: 100
+        }
+      }
+      momentum_optimizer_value: 0.9
+    }
+    use_moving_average: false
+  }
+  fine_tune_checkpoint: "../models/ssd_mobilenet_v2_320x320_coco17_tpu-8/checkpoint/ckpt-0"
+  num_steps: 3000
+  startup_delay_steps: 0.0
+  replicas_to_aggregate: 8
+  max_number_of_boxes: 100
+  unpad_groundtruth_tensors: false
+  fine_tune_checkpoint_type: "detection"
+  use_bfloat16: false
+  fine_tune_checkpoint_version: V2
+}
+train_input_reader {
+  label_map_path: "../models/raccoon_labelmap.pbtxt"
+  tf_record_input_reader {
+    input_path: "../data/raccoon_data/train.record"
+  }
+}
+eval_config {
+  metrics_set: "coco_detection_metrics"
+  use_moving_averages: false
+}
+eval_input_reader {
+  label_map_path: "../models/raccoon_labelmap.pbtxt"
+  shuffle: false
+  num_epochs: 1
+  tf_record_input_reader {
+    input_path: "../data/raccoon_data/val.record"
+  }
+}