The tensorflow object_detection model can only detect 1 face - tensorflow

Question

I'm trying to build a model to detect faces, and I used the SSD MobileNet v2 320x320's pipeline.config to train without calling the fine checkpoint. The pbtxt I used is the face_label_map.pbtxt that is given officially.
The model detects well when the photos only contain 1 face, but only detects one when there is more than 1 face in the photo.
I am searching for a long time on net. But no use. Please help or try to give some ideas on how to solve this. Thanks in advance.
result of one face and
result of more than one face
The code of pipeline.config:
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: 3.9999998989515007e-05
}
}
initializer {
truncated_normal_initializer {
mean: 0.0
stddev: 0.029999999329447746
}
}
activation: RELU_6
batch_norm {
decay: 0.9700000286102295
center: true
scale: true
epsilon: 0.0010000000474974513
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: 0.5
unmatched_threshold: 0.5
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: 3.9999998989515007e-05
}
}
initializer {
random_normal_initializer {
mean: 0.0
stddev: 0.009999999776482582
}
}
activation: RELU_6
batch_norm {
decay: 0.9700000286102295
center: true
scale: true
epsilon: 0.0010000000474974513
train: true
}
}
min_depth: 0
max_depth: 0
num_layers_before_predictor: 0
use_dropout: false
dropout_keep_probability: 0.800000011920929
kernel_size: 1
box_code_size: 4
apply_sigmoid_to_scores: false
class_prediction_bias_init: -4.599999904632568
}
}
anchor_generator {
ssd_anchor_generator {
num_layers: 6
min_scale: 0.20000000298023224
max_scale: 0.949999988079071
aspect_ratios: 1.0
aspect_ratios: 2.0
aspect_ratios: 0.5
aspect_ratios: 3.0
aspect_ratios: 0.33329999446868896
}
}
post_processing {
batch_non_max_suppression {
score_threshold: 9.99999993922529e-09
iou_threshold: 0.6000000238418579
max_detections_per_class: 100
max_total_detections: 100
use_static_shapes: false
}
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: 32
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: 0.800000011920929
total_steps: 50000
warmup_learning_rate: 0.13333000242710114
warmup_steps: 2000
}
}
momentum_optimizer_value: 0.8999999761581421
}
use_moving_average: false
}
num_steps: 50000
}
train_input_reader {
label_map_path: "/Users/aintor/Documents/Deep_Learning/Tensorflow/workspace/data/face_label_map.pbtxt"
tf_record_input_reader {
input_path: "/Users/aintor/Documents/Deep_Learning/Tensorflow/workspace/data/faces_train.tfrecord"
}
}
eval_config {
metrics_set: "coco_detection_metrics"
use_moving_averages: false
}
eval_input_reader {
label_map_path: "/Users/aintor/Documents/Deep_Learning/Tensorflow/workspace/data/face_label_map.pbtxt"
shuffle: false
num_epochs: 1
tf_record_input_reader {
input_path: "/Users/aintor/Documents/Deep_Learning/Tensorflow/workspace/data/faces_test.tfrecord"
}
}
The code of the pbtxt I used:
item {
name: "face"
id: 1
display_name: "face"
}
The code to generate the result photo:
import tensorflow as tf
import os
import numpy as np
from PIL import Image
import matplotlib
from matplotlib import pyplot as plt
import sys
sys.path.insert(0, path2scripts) # making scripts in models/research available for import
from object_detection.utils import label_map_util
from object_detection.utils import config_util
from object_detection.utils import visualization_utils as viz_utils
from object_detection.builders import model_builder
# do not change anything in this cell
configs = config_util.get_configs_from_pipeline_file(path2config) # importing config
model_config = configs['model'] # recreating model config
detection_model = model_builder.build(model_config=model_config, is_training=False) # importing model
ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
ckpt.restore(os.path.join(path2model, ckpt_num)).expect_partial()
category_index = label_map_util.create_category_index_from_labelmap(path2label_map,use_display_name=True)
def detect_fn(image):
image, shapes = detection_model.preprocess(image)
prediction_dict = detection_model.predict(image, shapes)
detections = detection_model.postprocess(prediction_dict, shapes)
return detections
def load_image_into_numpy_array(path):
"""Load an image from file into a numpy array.
Puts image into numpy array to feed into tensorflow graph.
Note that by convention we put it into a numpy array with shape
(height, width, channels), where channels=3 for RGB.
Args:
path: the file path to the image
Returns:
numpy array with shape (img_height, img_width, 3)
"""
return np.array(Image.open(path))
def inference_with_plot(path2images, box_th=0.25):
"""
Function that performs inference and plots resulting b-boxes
Args:
path2images: an array with pathes to images
box_th: (float) value that defines threshold for model prediction.
Returns:
None
"""
matplotlib.use('MacOSX')
for image_path in path2images:
print('Running inference for {}... '.format(image_path), end='')
image_np = load_image_into_numpy_array(image_path)
input_tensor = tf.convert_to_tensor(np.expand_dims(image_np, 0), dtype=tf.float32)
detections = detect_fn(input_tensor)
# All outputs are batches tensors.
# Convert to numpy arrays, and take index [0] to remove the batch dimension.
# We're only interested in the first num_detections.
num_detections = int(detections.pop('num_detections'))
detections = {key: value[0, :num_detections].numpy()
for key, value in detections.items()}
detections['num_detections'] = num_detections
# detection_classes should be ints.
detections['detection_classes'] = detections['detection_classes'].astype(np.int64)
label_id_offset = 1
image_np_with_detections = image_np.copy()
viz_utils.visualize_boxes_and_labels_on_image_array(
image_np_with_detections,
detections['detection_boxes'],
detections['detection_classes']+label_id_offset,
detections['detection_scores'],
category_index,
use_normalized_coordinates=True,
max_boxes_to_draw=200,
min_score_thresh=box_th,
agnostic_mode=False,
line_thickness=5)
plt.figure(figsize=(15,10))
plt.imshow(image_np_with_detections)
print('Done')
plt.show()
inference_with_plot(test_images)