Alex Chan
initial commit
999c5c9
"""
DeepLabCut Toolbox (deeplabcut.org)
© A. & M. Mathis Labs
Licensed under GNU Lesser General Public License v3.0
"""
import os
import ruamel.yaml
import glob
import warnings
import numpy as np
import tensorflow as tf
import typing
from pathlib import Path
from typing import Optional, Tuple, List
try:
TFVER = [int(v) for v in tf.__version__.split(".")]
if TFVER[1] < 14:
from tensorflow.contrib.tensorrt import trt_convert as trt
else:
from tensorflow.python.compiler.tensorrt import trt_convert as trt
except Exception:
pass
from dlclive.graph import (
read_graph,
finalize_graph,
get_output_nodes,
get_output_tensors,
extract_graph,
)
from dlclive.pose import extract_cnn_output, argmax_pose_predict, multi_pose_predict
from dlclive.display import Display
from dlclive import utils
from dlclive.exceptions import DLCLiveError, DLCLiveWarning
if typing.TYPE_CHECKING:
from dlclive.processor import Processor
class DLCLive(object):
"""
Object that loads a DLC network and performs inference on single images (e.g. images captured from a camera feed)
Parameters
-----------
path : string
Full path to exported model directory
model_type: string, optional
which model to use: 'base', 'tensorrt' for tensorrt optimized graph, 'lite' for tensorflow lite optimized graph
precision : string, optional
precision of model weights, only for model_type='tensorrt'. Can be 'FP16' (default), 'FP32', or 'INT8'
cropping : list of int
cropping parameters in pixel number: [x1, x2, y1, y2]
dynamic: triple containing (state, detectiontreshold, margin)
If the state is true, then dynamic cropping will be performed. That means that if an object is detected (i.e. any body part > detectiontreshold),
then object boundaries are computed according to the smallest/largest x position and smallest/largest y position of all body parts. This window is
expanded by the margin and from then on only the posture within this crop is analyzed (until the object is lost, i.e. <detectiontreshold). The
current position is utilized for updating the crop window for the next frame (this is why the margin is important and should be set large
enough given the movement of the animal).
resize : float, optional
Factor to resize the image.
For example, resize=0.5 will downsize both the height and width of the image by a factor of 2.
processor: dlc pose processor object, optional
User-defined processor object. Must contain two methods: process and save.
The 'process' method takes in a pose, performs some processing, and returns processed pose.
The 'save' method saves any valuable data created by or used by the processor
Processors can be used for two main purposes:
i) to run a forward predicting model that will predict the future pose from past history of poses (history can be stored in the processor object, but is not stored in this DLCLive object)
ii) to trigger external hardware based on pose estimation (e.g. see 'TeensyLaser' processor)
convert2rgb : bool, optional
boolean flag to convert frames from BGR to RGB color scheme
display : bool, optional
Display frames with DeepLabCut labels?
This is useful for testing model accuracy and cropping parameters, but it is very slow.
display_lik : float, optional
Likelihood threshold for display
display_raidus : int, optional
radius for keypoint display in pixels, default=3
"""
PARAMETERS = (
"path",
"cfg",
"model_type",
"precision",
"cropping",
"dynamic",
"resize",
"processor",
)
def __init__(
self,
model_path:str,
model_type:str="base",
precision:str="FP32",
tf_config=None,
cropping:Optional[List[int]]=None,
dynamic:Tuple[bool, float, float]=(False, 0.5, 10),
resize:Optional[float]=None,
convert2rgb:bool=True,
processor:Optional['Processor']=None,
display:typing.Union[bool, Display]=False,
pcutoff:float=0.5,
display_radius:int=3,
display_cmap:str="bmy",
):
self.path = model_path
self.cfg = None # type: typing.Optional[dict]
self.model_type = model_type
self.tf_config = tf_config
self.precision = precision
self.cropping = cropping
self.dynamic = dynamic
self.dynamic_cropping = None
self.resize = resize
self.processor = processor
self.convert2rgb = convert2rgb
if isinstance(display, Display):
self.display = display
elif display:
self.display = Display(pcutoff=pcutoff, radius=display_radius, cmap=display_cmap)
else:
self.display = None
self.sess = None
self.inputs = None
self.outputs = None
self.tflite_interpreter = None
self.pose = None
self.is_initialized = False
# checks
if self.model_type == "tflite" and self.dynamic[0]:
self.dynamic = (False, *self.dynamic[1:])
warnings.warn(
"Dynamic cropping is not supported for tensorflow lite inference. Dynamic cropping will not be used...",
DLCLiveWarning,
)
self.read_config()
def read_config(self):
""" Reads configuration yaml file
Raises
------
FileNotFoundError
error thrown if pose configuration file does nott exist
"""
cfg_path = Path(self.path).resolve() / "pose_cfg.yaml"
if not cfg_path.exists():
raise FileNotFoundError(
f"The pose configuration file for the exported model at {str(cfg_path)} was not found. Please check the path to the exported model directory"
)
ruamel_file = ruamel.yaml.YAML()
self.cfg = ruamel_file.load(open(str(cfg_path), "r"))
@property
def parameterization(self) -> dict:
"""
Return
Returns
-------
"""
return {param: getattr(self, param) for param in self.PARAMETERS}
def process_frame(self, frame):
"""
Crops an image according to the object's cropping and dynamic properties.
Parameters
-----------
frame :class:`numpy.ndarray`
image as a numpy array
Returns
----------
frame :class:`numpy.ndarray`
processed frame: convert type, crop, convert color
"""
if frame.dtype != np.uint8:
frame = utils.convert_to_ubyte(frame)
if self.cropping:
frame = frame[
self.cropping[2] : self.cropping[3], self.cropping[0] : self.cropping[1]
]
if self.dynamic[0]:
if self.pose is not None:
detected = self.pose[:, 2] > self.dynamic[1]
if np.any(detected):
x = self.pose[detected, 0]
y = self.pose[detected, 1]
x1 = int(max([0, int(np.amin(x)) - self.dynamic[2]]))
x2 = int(min([frame.shape[1], int(np.amax(x)) + self.dynamic[2]]))
y1 = int(max([0, int(np.amin(y)) - self.dynamic[2]]))
y2 = int(min([frame.shape[0], int(np.amax(y)) + self.dynamic[2]]))
self.dynamic_cropping = [x1, x2, y1, y2]
frame = frame[y1:y2, x1:x2]
else:
self.dynamic_cropping = None
if self.resize != 1:
frame = utils.resize_frame(frame, self.resize)
if self.convert2rgb:
frame = utils.img_to_rgb(frame)
return frame
def init_inference(self, frame=None, **kwargs):
"""
Load model and perform inference on first frame -- the first inference is usually very slow.
Parameters
-----------
frame :class:`numpy.ndarray`
image as a numpy array
Returns
--------
pose :class:`numpy.ndarray`
the pose estimated by DeepLabCut for the input image
"""
# get model file
model_file = glob.glob(os.path.normpath(self.path + "/*.pb"))[0]
if not os.path.isfile(model_file):
raise FileNotFoundError(
"The model file {} does not exist.".format(model_file)
)
# process frame
if frame is None and (self.model_type == "tflite"):
raise DLCLiveError(
"No image was passed to initialize inference. An image must be passed to the init_inference method"
)
if frame is not None:
if frame.ndim == 2:
self.convert2rgb = True
processed_frame = self.process_frame(frame)
# load model
if self.model_type == "base":
graph_def = read_graph(model_file)
graph = finalize_graph(graph_def)
self.sess, self.inputs, self.outputs = extract_graph(
graph, tf_config=self.tf_config
)
elif self.model_type == "tflite":
###
# the frame size needed to initialize the tflite model as
# tflite does not support saving a model with dynamic input size
###
# get input and output tensor names from graph_def
graph_def = read_graph(model_file)
graph = finalize_graph(graph_def)
output_nodes = get_output_nodes(graph)
output_nodes = [on.replace("DLC/", "") for on in output_nodes]
tf_version_2 = tf.__version__[0] == '2'
if tf_version_2:
converter = tf.compat.v1.lite.TFLiteConverter.from_frozen_graph(
model_file,
["Placeholder"],
output_nodes,
input_shapes={"Placeholder": [1, processed_frame.shape[0], processed_frame.shape[1], 3]},
)
else:
converter = tf.lite.TFLiteConverter.from_frozen_graph(
model_file,
["Placeholder"],
output_nodes,
input_shapes={"Placeholder": [1, processed_frame.shape[0], processed_frame.shape[1], 3]},
)
try:
tflite_model = converter.convert()
except Exception:
raise DLCLiveError(
(
"This model cannot be converted to tensorflow lite format. "
"To use tensorflow lite for live inference, "
"make sure to set TFGPUinference=False "
"when exporting the model from DeepLabCut"
)
)
self.tflite_interpreter = tf.lite.Interpreter(model_content=tflite_model)
self.tflite_interpreter.allocate_tensors()
self.inputs = self.tflite_interpreter.get_input_details()
self.outputs = self.tflite_interpreter.get_output_details()
elif self.model_type == "tensorrt":
graph_def = read_graph(model_file)
graph = finalize_graph(graph_def)
output_tensors = get_output_tensors(graph)
output_tensors = [ot.replace("DLC/", "") for ot in output_tensors]
if (TFVER[0] > 1) | (TFVER[0] == 1 & TFVER[1] >= 14):
converter = trt.TrtGraphConverter(
input_graph_def=graph_def,
nodes_blacklist=output_tensors,
is_dynamic_op=True,
)
graph_def = converter.convert()
else:
graph_def = trt.create_inference_graph(
input_graph_def=graph_def,
outputs=output_tensors,
max_batch_size=1,
precision_mode=self.precision,
is_dynamic_op=True,
)
graph = finalize_graph(graph_def)
self.sess, self.inputs, self.outputs = extract_graph(
graph, tf_config=self.tf_config
)
else:
raise DLCLiveError(
"model_type = {} is not supported. model_type must be 'base', 'tflite', or 'tensorrt'".format(
self.model_type
)
)
# get pose of first frame (first inference is often very slow)
if frame is not None:
pose = self.get_pose(frame, **kwargs)
else:
pose = None
self.is_initialized = True
return pose
def get_pose(self, frame=None, **kwargs):
"""
Get the pose of an image
Parameters
-----------
frame :class:`numpy.ndarray`
image as a numpy array
Returns
--------
pose :class:`numpy.ndarray`
the pose estimated by DeepLabCut for the input image
"""
if frame is None:
raise DLCLiveError("No frame provided for live pose estimation")
frame = self.process_frame(frame)
if self.model_type in ["base", "tensorrt"]:
pose_output = self.sess.run(
self.outputs, feed_dict={self.inputs: np.expand_dims(frame, axis=0)}
)
elif self.model_type == "tflite":
self.tflite_interpreter.set_tensor(
self.inputs[0]["index"],
np.expand_dims(frame, axis=0).astype(np.float32),
)
self.tflite_interpreter.invoke()
if len(self.outputs) > 1:
pose_output = [
self.tflite_interpreter.get_tensor(self.outputs[0]["index"]),
self.tflite_interpreter.get_tensor(self.outputs[1]["index"]),
]
else:
pose_output = self.tflite_interpreter.get_tensor(
self.outputs[0]["index"]
)
else:
raise DLCLiveError(
"model_type = {} is not supported. model_type must be 'base', 'tflite', or 'tensorrt'".format(
self.model_type
)
)
# check if using TFGPUinference flag
# if not, get pose from network output
if len(pose_output) > 1:
scmap, locref = extract_cnn_output(pose_output, self.cfg)
num_outputs = self.cfg.get("num_outputs", 1)
if num_outputs > 1:
self.pose = multi_pose_predict(
scmap, locref, self.cfg["stride"], num_outputs
)
else:
self.pose = argmax_pose_predict(scmap, locref, self.cfg["stride"])
else:
pose = np.array(pose_output[0])
self.pose = pose[:, [1, 0, 2]]
# display image if display=True before correcting pose for cropping/resizing
if self.display is not None:
self.display.display_frame(frame, self.pose)
# if frame is cropped, convert pose coordinates to original frame coordinates
if self.resize is not None:
self.pose[:, :2] *= 1 / self.resize
if self.cropping is not None:
self.pose[:, 0] += self.cropping[0]
self.pose[:, 1] += self.cropping[2]
if self.dynamic_cropping is not None:
self.pose[:, 0] += self.dynamic_cropping[0]
self.pose[:, 1] += self.dynamic_cropping[2]
# process the pose
if self.processor:
self.pose = self.processor.process(self.pose, **kwargs)
return self.pose
def close(self):
""" Close tensorflow session
"""
self.sess.close()
self.sess = None
self.is_initialized = False
if self.display is not None:
self.display.destroy()