hailo-inference/inference.py
raphael be89e739b4 refactors into classes
- ImageMeta struct saves image metadata
- DataHandler takes care of loading dataset
  and parsing the label information
- YoloProcessing takes care of image preprocessing
  and yolo postprocessing
- HailoHandler connects to Hailo device, pushes
  the desired network hef file and runs the dataset
  on the hailo chip
2022-03-02 11:44:50 +01:00

314 lines
13 KiB
Python

import json
import os
import time
t
from PIL import Image
from detection_tools.utils.visualization_utils import \
visualize_boxes_and_labels_on_image_array
from hailo_platform import (ConfigureParams, FormatType, HEF,
HailoStreamInterface, InferVStreams,
InputVStreamParams, OutputVStreamParams,
PcieDevice)
import numpy as np
import tensorflow as tf
from tensorflow.image import combined_non_max_suppression
# Collect images from data files
class ImageMeta:
def __init__(self, image_height, image_width, channels):
self.image_height = image_height
self.image_width = image_width
self.channels = channels
class DataHandler:
def __init__(self, path, image_meta):
self.images_path = path
self.image_meta = image_meta
def load_data(self, preprocess_fn):
names = []
images_list = [img_name for img_name in os.listdir(self.images_path)
if os.path.splitext(os.path.join(self.images_path, img_name))[1] == '.jpg']
dataset = np.zeros((1, self.image_meta.image_height,
self.image_meta.image_width,
self.image_meta.channels),
dtype=np.float32)
for idx, img_name in enumerate(images_list):
img = Image.open(os.path.join(self.images_path, img_name))
img_preproc = preprocess_fn(img)
dataset[idx, :, :, :] = np.array(img_preproc)
names.append(img_name)
break
self.dataset = dataset
self.names = names
def _get_coco_labels(self):
coco_names = json.load(open(os.path.join(os.path.dirname(__file__), 'coco_names.json')))
coco_names = {int(k): {'id': int(k), 'name': str(v)} for (k, v) in coco_names.items()}
return coco_names
def _get_labels(self, label_name):
filename = os.path.join(os.path.dirname(__file__), label_name + '.json')
names = json.load(open(filename))
names = {int(k): {'id': int(k), 'name': str(v)} for (k, v) in names.items()}
return names
COCO_17_14 = {0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5, 6: 6, 7: 7, 8: 8, 9: 9,
10: 10, 11: 11, 12: 13, 13: 14, 14: 15, 15: 16, 16: 17, 17: 18,
18: 19, 19: 20, 20: 21, 21: 22, 22: 23, 23: 24, 24: 25, 25: 27,
26: 28, 27: 31, 28: 32, 29: 33, 30: 34, 31: 35, 32: 36, 33: 37,
34: 38, 35: 39, 36: 40, 37: 41, 38: 42, 39: 43, 40: 44, 41: 46,
42: 47, 43: 48, 44: 49, 45: 50, 46: 51, 47: 52, 48: 53, 49: 54,
50: 55, 51: 56, 52: 57, 53: 58, 54: 59, 55: 60, 56: 61, 57: 62,
58: 63, 59: 64, 60: 65, 61: 67, 62: 70, 63: 72, 64: 73, 65: 74,
66: 75, 67: 76, 68: 77, 69: 78, 70: 79, 71: 80, 72: 81, 73: 82,
74: 84, 75: 85, 76: 86, 77: 87, 78: 88, 79: 89, 80: 90}
class YoloProcessing:
def __init__(self, imageMeta, classes):
self.output_height = imageMeta.image_height
self.output_width = imageMeta.image_width
self.classes = classes
def preproc(self, image, resize_side=256):
'''
imagenet-standard: aspect-preserving resize to 256px smaller-side,
then central-crop to 224px
'''
new_width = int(image.width/image.height*resize_side)
new_height = resize_side
x, y = (new_width-self.output_width)/2, 0
# Select area to crop
area = (x, y, x+self.output_width, y+self.output_height)
# Crop, show, and save image
cropped_img = image.resize((new_width, new_height)).crop(area)
return cropped_img
# 20 x 20 -> 32
# stride = 32
def yolo_postprocess_numpy(self, net_out, anchors_for_stride, stride):
"""
net_out is shape: [N, 19, 19, 255] or [N, 38, 38, 255] or [N, 76, 76, 255]
first we reshape it to be as in gluon and then follow gluon's shapes.
output_ind = 0 for stride 32, 1 for stride 16, 2 for stride 8.
"""
# net_out = net_out.astype(np.float32) / 256
num_classes = 4
BS = net_out.shape[0] # batch size
H = net_out.shape[1]
W = net_out.shape[2]
num_anchors = anchors_for_stride.size // 2 # 2 params for each anchor.
num_pred = 1 + 4 + num_classes # 2 box centers, 2 box scales, 1 objness, num_classes class scores
alloc_size = (128, 128)
grid_x = np.arange(alloc_size[1])
grid_y = np.arange(alloc_size[0])
grid_x, grid_y = np.meshgrid(grid_x, grid_y) # dims [128,128], [128,128]
offsets = np.concatenate((grid_x[:, :, np.newaxis], grid_y[:, :, np.newaxis]), axis=-1) # dim [128,128,2]
offsets = np.expand_dims(np.expand_dims(offsets, 0), 0) # dim [1,1,128,128,2]
pred = net_out.transpose((0, 3, 1, 2)) # now dims are: [N,C,H,W] as in Gluon.
pred = np.reshape(pred, (BS, num_anchors * num_pred, -1)) # dim [N, 255, HxW]
# dim [N, 361, 255], we did it so that the 255 be the last dim and can be reshaped.
pred = pred.transpose((0, 2, 1))
pred = np.reshape(pred, (BS, -1, num_anchors, num_pred)) # dim [N, HxW, 3, 85]]
raw_box_centers = pred[:, :, :, 0:2] # dim [N, HxW, 3, 2]
raw_box_scales = pred[:, :, :, 2:4] # dim [N,HxW, 3, 2]
objness = pred[:, :, :, 4:5] # dim [N, HxW, 3, 1]
class_pred = pred[:, :, :, 5:] # dim [N, HxW, 3, 80]
offsets = offsets[:, :, :H, :W, :] # dim [1, 1, H, W, 2]
offsets = np.reshape(offsets, (1, -1, 1, 2)) # dim [1, HxW, 1, 2]
box_centers, box_scales, confidence, class_pred = self._yolo5_decode(
raw_box_centers=raw_box_centers,
raw_box_scales=raw_box_scales,
objness=objness,
class_pred=class_pred,
anchors_for_stride=anchors_for_stride,
offsets=offsets,
stride=stride)
class_score = class_pred * confidence # dim [N, HxW, 3, 80]
wh = box_scales / 2.0
# dim [N, HxW, 3, 4]. scheme xmin, ymin, xmax, ymax
bbox = np.concatenate((box_centers - wh, box_centers + wh), axis=-1)
detection_boxes = np.reshape(bbox, (BS, -1, 1, 4)) # dim [N, num_detections, 1, 4]
detection_scores = np.reshape(class_score, (BS, -1, num_classes)) # dim [N, num_detections, 80]
# switching scheme from xmin, ymin, xmanx, ymax to ymin, xmin, ymax, xmax:
detection_boxes_tmp = np.zeros(detection_boxes.shape)
detection_boxes_tmp[:, :, :, 0] = detection_boxes[:, :, :, 1]
detection_boxes_tmp[:, :, :, 1] = detection_boxes[:, :, :, 0]
detection_boxes_tmp[:, :, :, 2] = detection_boxes[:, :, :, 3]
detection_boxes_tmp[:, :, :, 3] = detection_boxes[:, :, :, 2]
detection_boxes = detection_boxes_tmp # now scheme is: ymin, xmin, ymax, xmax
return detection_boxes.astype(np.float32), detection_scores.astype(np.float32)
def _yolo5_decode(self, raw_box_centers, raw_box_scales, objness, class_pred, anchors_for_stride, offsets, stride):
box_centers = (raw_box_centers * 2. - 0.5 + offsets) * stride
box_scales = (raw_box_scales * 2) ** 2 * anchors_for_stride # dim [N, HxW, 3, 2]
return box_centers, box_scales, objness, class_pred
def postprocessing(self, endnodes):
"""
endnodes is a list of 3 output tensors:
endnodes[0] - stride 32 of input
endnodes[1] - stride 16 of input
endnodes[2] - stride 8 of input
Returns:
a tensor with dims: [BS, Total_num_of_detections_in_image, 6]
where:
total_num_of_detections_in_image = H*W*((1/32^2) + (1/16^2) + (1/8^2))*num_anchors*num_classes,
with H, W as input dims.
If H=W=608, num_anchors=3, num_classes=80 (coco 2017), we get:
total_num_of_detections = 1819440 ~ 1.8M detections per image for the NMS
"""
H_input = 640
W_input = 640
anchors_list = [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]]
# TODO make prettier
strides = [8, 16, 32]
for output_ind, output_branch in enumerate(endnodes): # iterating over the output layers:
stride = strides[::-1][output_ind]
anchors_for_stride = np.array(anchors_list[::-1][output_ind])
anchors_for_stride = np.reshape(anchors_for_stride, (1, 1, -1, 2)) # dim [1, 1, 3, 2]
detection_boxes, detection_scores = self.yolo_postprocess_numpy(output_branch,
anchors_for_stride,
stride)
# detection_boxes is a [BS, num_detections, 1, 4] tensor, detection_scores is a
# [BS, num_detections, num_classes] tensor
detection_boxes = detection_boxes / H_input # normalization of box coordinates to 1
BS = endnodes[0].shape[0]
H = H_input // stride
W = W_input // stride
num_anchors = anchors_for_stride.size // 2
num_detections = H * W * num_anchors
# detection_boxes.set_shape((BS, num_detections, 1, 4))
# detection_scores.set_shape((BS, num_detections, num_classes))
# concatenating the detections from the different output layers:
if output_ind == 0:
detection_boxes_full = detection_boxes
detection_scores_full = detection_scores
else:
detection_boxes_full = tf.concat([detection_boxes_full, detection_boxes], axis=1)
detection_scores_full = tf.concat([detection_scores_full, detection_scores], axis=1)
score_threshold = 0.5
nms_iou_threshold = 0.5
labels_offset = 1
(nmsed_boxes, nmsed_scores, nmsed_classes, num_detections) = \
combined_non_max_suppression(boxes=detection_boxes_full,
scores=detection_scores_full,
score_threshold=score_threshold,
iou_threshold=nms_iou_threshold,
max_output_size_per_class=100,
max_total_size=100)
# adding offset to the class prediction and cast to integer
def translate_coco_2017_to_2014(nmsed_classes):
return np.vectorize(COCO_17_14.get)(nmsed_classes).astype(np.int32)
nmsed_classes = tf.cast(tf.add(nmsed_classes, labels_offset), tf.int16)
nmsed_classes = translate_coco_2017_to_2014(nmsed_classes)
return {'detection_boxes': nmsed_boxes,
'detection_scores': nmsed_scores,
'detection_classes': nmsed_classes,
'num_detections': num_detections}
class HailoHandler:
def __init__(self, hef_path='hef/yolov5m.hef'):
target = PcieDevice()
self.hef = HEF(hef_path)
# Configure network groups
configure_params = ConfigureParams.create_from_hef(hef=self.hef,
interface=HailoStreamInterface.PCIe)
network_groups = target.configure(self.hef, configure_params)
self.network_group = network_groups[0]
self.input_vstreams_params = InputVStreamParams.make(self.network_group,
quantized=False,
format_type=FormatType.FLOAT32)
self.output_vstreams_params = OutputVStreamParams.make(self.network_group, quantized=False, format_type=FormatType.FLOAT32)
self.input_vstream_info = self.hef.get_input_vstream_infos()[0]
self.output_vstream_infos = self.hef.get_output_vstream_infos()
self.network_group_params = self.network_group.create_params()
def run_hailo(self, dataset):
input_data = {self.input_vstream_info.name: dataset}
with InferVStreams(self.network_group, self.input_vstreams_params, self.output_vstreams_params) as infer_pipeline:
with self.network_group.activate(self.network_group_params):
infer_results = infer_pipeline.infer(input_data)
out = [infer_results[i.name] for i in self.output_vstream_infos]
return out
def process_yolo5():
imageMeta = ImageMeta(640, 640, 3)
processor = YoloProcessing(imageMeta, classes=3)
data = DataHandler('./minimal_data', imageMeta)
data.load_data(processor.preproc)
hailo = HailoHandler('hef/yolov5m_22_2.hef')
out = hailo.run_hailo(data.dataset)
logits = processor.postprocessing(out)
labels = data._get_labels("daria_names")
image = visualize_boxes_and_labels_on_image_array(
data.dataset[0],
logits['detection_boxes'].numpy()[0],
logits['detection_classes'][0],
logits['detection_scores'].numpy()[0],
labels,
use_normalized_coordinates=True,
max_boxes_to_draw=100,
min_score_thresh=.5,
agnostic_mode=False,
line_thickness=4)
Image.fromarray(np.uint8(image)).save('/home/maintenance/test.png')
if __name__ == "__main__":
process_yolo5()