Untitled
unknown
python
2 years ago
20 kB
6
Indexable
#!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
import numpy as np
import argparse
import glob
import logging
import os
import pickle
import sys
from typing import Any, ClassVar, Dict, List
import torch
import cv2
from detectron2.config import CfgNode, get_cfg
from detectron2.data.detection_utils import read_image
from detectron2.engine.defaults import DefaultPredictor
from detectron2.structures.instances import Instances
from detectron2.utils.logger import setup_logger
from densepose import add_densepose_config
from densepose.structures import DensePoseChartPredictorOutput, DensePoseEmbeddingPredictorOutput
from densepose.utils.logger import verbosity_to_level
from densepose.vis.base import CompoundVisualizer
from densepose.vis.bounding_box import ScoredBoundingBoxVisualizer
from densepose.vis.densepose_outputs_vertex import (
DensePoseOutputsTextureVisualizer,
DensePoseOutputsVertexVisualizer,
get_texture_atlases,
)
from densepose.vis.densepose_results import (
DensePoseResultsContourVisualizer,
DensePoseResultsFineSegmentationVisualizer,
DensePoseResultsUVisualizer,
DensePoseResultsVVisualizer,
)
from densepose.vis.densepose_results_textures import (
DensePoseResultsVisualizerWithTexture,
get_texture_atlas,
)
from densepose.vis.extractor import (
CompoundExtractor,
DensePoseOutputsExtractor,
DensePoseResultExtractor,
create_extractor,
)
from pdb import set_trace as bb
import tqdm
DOC = """Apply Net - a tool to print / visualize DensePose results
"""
LOGGER_NAME = "apply_net"
logger = logging.getLogger(LOGGER_NAME)
_ACTION_REGISTRY: Dict[str, "Action"] = {}
def box_overlaps(box1, box2):
x1_min, y1_min, x1_max, y1_max = box1
x2_min, y2_min, x2_max, y2_max = box2
# Compute the area of both bounding boxes
area1 = (x1_max - x1_min) * (y1_max - y1_min)
area2 = (x2_max - x2_min) * (y2_max - y2_min)
# Compute the intersection area
x_min = max(x1_min, x2_min)
y_min = max(y1_min, y2_min)
x_max = min(x1_max, x2_max)
y_max = min(y1_max, y2_max)
intersection_area = max(0, x_max - x_min) * max(0, y_max - y_min)
# Compute the union area
union_area = area1 + area2 - intersection_area
# Compute the IoU
iou = intersection_area / union_area
return iou
def get_box(mask):
if np.sum(mask) == 0:
return -1,-1
ret,binary = cv2.threshold(mask,127,255,cv2.THRESH_BINARY)
contours,hierarchy = cv2.findContours(binary,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)
x,y,w,h = cv2.boundingRect(contours[0])
center = (x + w//2, y + h//2)
if max(w, h) < 80 or max(w,h)/max(mask.shape[0], mask.shape[1]) < 0.05:
return -1, -1
d = max(w,h)//2
if d < 74:
d = 74
y2 = min(center[1] + 1.5*d, mask.shape[0]-1)
y1 = y2 - 3.5*d
if y1 < 0:
y1 = 0
y2 = y1 + 3.5*d
x1 = max(0, center[0] - 1.75*d)
x2 = x1 + d*3.5
if x2 >= mask.shape[1]:
x2 = mask.shape[1]-1
x1 = x2 - d*3.5
return [int(x1),int(y1),int(x2),int(y2)], [int(x),int(y),int(x+w),int(y+h)]
def get_arm_box(img, mask):
if np.sum(mask) == 0:
return -1,-1
bb()
skin_mask = get_skin_color_mask(img.copy())
skin_mask[mask==0] = 0
skin_pixels = np.sum(skin_mask>0)
arm_pixels = np.sum(mask>0)
ratio = skin_pixels/arm_pixels
if ratio < 0.3:
return -1,-1
ret,binary = cv2.threshold(mask,127,255,cv2.THRESH_BINARY)
contours,hierarchy = cv2.findContours(binary,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)
x,y,w,h = cv2.boundingRect(contours[0])
center = (x + w//2, y + h//2)
if h < 500:
return -1, -1
if h/w < 2:
return -1, -1
d = h//2
y2 = min(center[1] + 1.05*d, mask.shape[0]-1)
y1 = y2 - 2.05*d
if y1 < 0:
y1 = 0
y2 = y1 + 2.05*d
x1 = max(0, center[0] - d*0.25)
x2 = x1 + d*0.5
if x2 >= mask.shape[1]:
x2 = mask.shape[1]-1
x1 = x2 - d*0.5
return [int(x1),int(y1),int(x2),int(y2)], [int(x),int(y),int(x+w),int(y+h)]
class Action(object):
@classmethod
def add_arguments(cls: type, parser: argparse.ArgumentParser):
parser.add_argument(
"-v",
"--verbosity",
action="count",
help="Verbose mode. Multiple -v options increase the verbosity.",
)
def register_action(cls: type):
"""
Decorator for action classes to automate action registration
"""
global _ACTION_REGISTRY
_ACTION_REGISTRY[cls.COMMAND] = cls
return cls
class InferenceAction(Action):
@classmethod
def add_arguments(cls: type, parser: argparse.ArgumentParser):
super(InferenceAction, cls).add_arguments(parser)
parser.add_argument("cfg", metavar="<config>", help="Config file")
parser.add_argument("model", metavar="<model>", help="Model file")
parser.add_argument("input", metavar="<input>", help="Input data")
parser.add_argument(
"--opts",
help="Modify config options using the command-line 'KEY VALUE' pairs",
default=[],
nargs=argparse.REMAINDER,
)
@classmethod
def execute(cls: type, args: argparse.Namespace):
logger.info(f"Loading config from {args.cfg}")
opts = []
cfg = cls.setup_config(args.cfg, args.model, args, opts)
logger.info(f"Loading model from {args.model}")
predictor = DefaultPredictor(cfg)
logger.info(f"Loading data from {args.input}")
file_list = cls._get_input_file_list(args.input)
if len(file_list) == 0:
logger.warning(f"No input images for {args.input}")
return
context = cls.create_context(args, cfg)
f = open(file_list[0], 'r')
file_list = f.read().split()
f.close()
num_images = len(file_list)
import time
np.random.seed(int(time.time()))
perm = np.random.permutation(len(file_list))
cfg['INPUT']['MIN_SIZE_TEST'] = 400
cfg['INPUT']['MAX_SIZE_TEST'] = 1200
#done_names = os.popen('ls /media/bharat/ssd/hand_data_2/*.jpg').read().split()
for i in tqdm.tqdm(range(len(file_list))):
file_name = '/media/bharat/ssd/downloads/fashionova/zalando/' + file_list[perm[i]]
name = file_name.split('.jpg')[0]
name = name + '.txt'
if os.path.exists(name):
continue
try:
img = read_image(file_name, format="BGR") # predictor expects BGR image.
except:
continue
print(name)
with torch.no_grad():
outputs = predictor(img)["instances"]
cls.execute_on_outputs(context, {"file_name": file_name, "image": img}, outputs)
os.system('touch ' + name)
cls.postexecute(context)
@classmethod
def setup_config(
cls: type, config_fpath: str, model_fpath: str, args: argparse.Namespace, opts: List[str]
):
cfg = get_cfg()
add_densepose_config(cfg)
cfg.merge_from_file(config_fpath)
cfg.merge_from_list(args.opts)
if opts:
cfg.merge_from_list(opts)
cfg.MODEL.WEIGHTS = model_fpath
cfg.freeze()
return cfg
@classmethod
def _get_input_file_list(cls: type, input_spec: str):
if os.path.isdir(input_spec):
file_list = [
os.path.join(input_spec, fname)
for fname in os.listdir(input_spec)
if os.path.isfile(os.path.join(input_spec, fname))
]
elif os.path.isfile(input_spec):
file_list = [input_spec]
else:
file_list = glob.glob(input_spec)
ffile_list = []
for names in file_list:
if '.png' in names:
continue
else:
ffile_list.append(names)
return ffile_list
@register_action
class DumpAction(InferenceAction):
"""
Dump action that outputs results to a pickle file
"""
COMMAND: ClassVar[str] = "dump"
@classmethod
def add_parser(cls: type, subparsers: argparse._SubParsersAction):
parser = subparsers.add_parser(cls.COMMAND, help="Dump model outputs to a file.")
cls.add_arguments(parser)
parser.set_defaults(func=cls.execute)
@classmethod
def add_arguments(cls: type, parser: argparse.ArgumentParser):
super(DumpAction, cls).add_arguments(parser)
parser.add_argument(
"--output",
metavar="<dump_file>",
default="results.pkl",
help="File name to save dump to",
)
@classmethod
def execute_on_outputs(
cls: type, context: Dict[str, Any], entry: Dict[str, Any], outputs: Instances
):
image_fpath = entry["file_name"]
logger.info(f"Processing {image_fpath}")
result = {"file_name": image_fpath}
if outputs.has("scores"):
result["scores"] = outputs.get("scores").cpu()
if outputs.has("pred_boxes"):
result["pred_boxes_XYXY"] = outputs.get("pred_boxes").tensor.cpu()
if outputs.has("pred_densepose"):
if isinstance(outputs.pred_densepose, DensePoseChartPredictorOutput):
extractor = DensePoseResultExtractor()
elif isinstance(outputs.pred_densepose, DensePoseEmbeddingPredictorOutput):
extractor = DensePoseOutputsExtractor()
result["pred_densepose"] = extractor(outputs)[0]
context["results"].append(result)
@classmethod
def create_context(cls: type, args: argparse.Namespace, cfg: CfgNode):
context = {"results": [], "out_fname": args.output}
return context
@classmethod
def postexecute(cls: type, context: Dict[str, Any]):
out_fname = context["out_fname"]
out_dir = os.path.dirname(out_fname)
if len(out_dir) > 0 and not os.path.exists(out_dir):
os.makedirs(out_dir)
with open(out_fname, "wb") as hFile:
pickle.dump(context["results"], hFile)
logger.info(f"Output saved to {out_fname}")
@register_action
class ShowAction(InferenceAction):
"""
Show action that visualizes selected entries on an image
"""
COMMAND: ClassVar[str] = "show"
VISUALIZERS: ClassVar[Dict[str, object]] = {
"dp_contour": DensePoseResultsContourVisualizer,
"dp_segm": DensePoseResultsFineSegmentationVisualizer,
"dp_u": DensePoseResultsUVisualizer,
"dp_v": DensePoseResultsVVisualizer,
"dp_iuv_texture": DensePoseResultsVisualizerWithTexture,
"dp_cse_texture": DensePoseOutputsTextureVisualizer,
"dp_vertex": DensePoseOutputsVertexVisualizer,
"bbox": ScoredBoundingBoxVisualizer,
}
@classmethod
def add_parser(cls: type, subparsers: argparse._SubParsersAction):
parser = subparsers.add_parser(cls.COMMAND, help="Visualize selected entries")
cls.add_arguments(parser)
parser.set_defaults(func=cls.execute)
@classmethod
def add_arguments(cls: type, parser: argparse.ArgumentParser):
super(ShowAction, cls).add_arguments(parser)
parser.add_argument(
"visualizations",
metavar="<visualizations>",
help="Comma separated list of visualizations, possible values: "
"[{}]".format(",".join(sorted(cls.VISUALIZERS.keys()))),
)
parser.add_argument(
"--min_score",
metavar="<score>",
default=0.9,
type=float,
help="Minimum detection score to visualize",
)
parser.add_argument(
"--nms_thresh", metavar="<threshold>", default=None, type=float, help="NMS threshold"
)
parser.add_argument(
"--texture_atlas",
metavar="<texture_atlas>",
default=None,
help="Texture atlas file (for IUV texture transfer)",
)
parser.add_argument(
"--texture_atlases_map",
metavar="<texture_atlases_map>",
default=None,
help="JSON string of a dict containing texture atlas files for each mesh",
)
parser.add_argument(
"--output",
metavar="<image_file>",
default="outputres.png",
help="File name to save output to",
)
@classmethod
def setup_config(
cls: type, config_fpath: str, model_fpath: str, args: argparse.Namespace, opts: List[str]
):
opts.append("MODEL.ROI_HEADS.SCORE_THRESH_TEST")
opts.append(str(args.min_score))
if args.nms_thresh is not None:
opts.append("MODEL.ROI_HEADS.NMS_THRESH_TEST")
opts.append(str(args.nms_thresh))
cfg = super(ShowAction, cls).setup_config(config_fpath, model_fpath, args, opts)
return cfg
@classmethod
def execute_on_outputs(
cls: type, context: Dict[str, Any], entry: Dict[str, Any], outputs: Instances
):
import cv2
import numpy as np
visualizer = context["visualizer"]
extractor = context["extractor"]
image_fpath = entry["file_name"]
#mask_fpath = entry["file_name"].split('.jpeg')[0]
logger.info(f"Processing {image_fpath}")
image = cv2.cvtColor(entry["image"], cv2.COLOR_BGR2GRAY)
image = np.tile(image[:, :, np.newaxis], [1, 1, 3])
data = extractor(outputs)
scaling_x = image.shape[0]
scaling_y = image.shape[1]
if data[0][0] is not None:
try:
all_boxes = []
all_fin_boxes = []
all_masks = []
img = cv2.imread(entry["file_name"])
for i in range(len(data[0][0])):
box = data[1][0][i].cpu().numpy()
mask = data[0][0][i].labels.cpu().numpy()
#rmask = np.array(255*(mask==3), dtype=np.uint8)
rmask = np.array(255*((mask==3) | (mask==16) | (mask==18) | (mask==20) | (mask==22)), dtype=np.uint8)
right_hand_mask = np.zeros((image.shape[0], image.shape[1]), dtype=np.uint8)
right_hand_mask[int(box[1]):int(box[1])+rmask.shape[0], int(box[0]):int(box[0])+rmask.shape[1]] = rmask
#discard very small boxes and resize very large boxes to a reasonable size
right_box_fin, original_right_box = get_arm_box(img, right_hand_mask)
#lmask = np.array(255*(mask==4), dtype=np.uint8)
lmask = np.array(255*((mask==4) | (mask==19) | (mask==21) | (mask==15) | (mask==17)), dtype=np.uint8)
left_hand_mask = np.zeros((image.shape[0], image.shape[1]), dtype=np.uint8)
left_hand_mask[int(box[1]):int(box[1])+lmask.shape[0], int(box[0]):int(box[0])+lmask.shape[1]] = lmask
left_box_fin, original_left_box = get_arm_box(img, left_hand_mask)
#check if there is any overlap across boxes, if yes, discard overlapping boxes
if right_box_fin != -1:
all_boxes.append(original_right_box)
all_fin_boxes.append(right_box_fin)
all_masks.append(right_hand_mask)
if left_box_fin != -1:
all_boxes.append(original_left_box)
all_masks.append(left_hand_mask)
all_fin_boxes.append(left_box_fin)
for i in range(len(all_boxes)):
#save the image, box and the mask
img = cv2.imread(entry["file_name"])
crop_box = all_fin_boxes[i]
img = img[crop_box[1]:crop_box[3], crop_box[0]:crop_box[2], :]
mask = all_masks[i][crop_box[1]:crop_box[3], crop_box[0]:crop_box[2]]
bbox = [all_boxes[i][0]-crop_box[0], all_boxes[i][1]-crop_box[1], all_boxes[i][2]-crop_box[0], all_boxes[i][3]-crop_box[1]]
name = entry["file_name"].split('.jpg')[0]
init_path = entry["file_name"].split('/')
init_path = init_path[-3] + '/' + init_path[-2] + '/'
base_name = '/media/bharat/ssd/arm_data/' + init_path
if not os.path.exists(base_name):
os.system('mkdir -p ' + base_name)
name = base_name + name.split('/')[-1]
#name = 'tmp2/' + name.split('/')[-1]
cv2.imwrite(name + '_{}_{}_{}_{}_{}.jpg'.format(i, bbox[0], bbox[1], bbox[2], bbox[3]), img)
cv2.imwrite(name + '_{}_mask.png'.format(i), mask)
except:
return
"""image_vis = visualizer.visualize(image, data)
entry_idx = context["entry_idx"] + 1
out_fname = cls._get_out_fname(entry_idx, context["out_fname"])
out_dir = os.path.dirname(out_fname)
if len(out_dir) > 0 and not os.path.exists(out_dir):
os.makedirs(out_dir)
#cv2.imwrite(out_fname, image_vis)
logger.info(f"Output saved to {out_fname}")
context["entry_idx"] += 1"""
@classmethod
def postexecute(cls: type, context: Dict[str, Any]):
pass
@classmethod
def _get_out_fname(cls: type, entry_idx: int, fname_base: str):
base, ext = os.path.splitext(fname_base)
return base + ".{0:04d}".format(entry_idx) + ext
@classmethod
def create_context(cls: type, args: argparse.Namespace, cfg: CfgNode) -> Dict[str, Any]:
vis_specs = args.visualizations.split(",")
visualizers = []
extractors = []
for vis_spec in vis_specs:
texture_atlas = get_texture_atlas(args.texture_atlas)
texture_atlases_dict = get_texture_atlases(args.texture_atlases_map)
vis = cls.VISUALIZERS[vis_spec](
cfg=cfg,
texture_atlas=texture_atlas,
texture_atlases_dict=texture_atlases_dict,
)
visualizers.append(vis)
extractor = create_extractor(vis)
extractors.append(extractor)
visualizer = CompoundVisualizer(visualizers)
extractor = CompoundExtractor(extractors)
context = {
"extractor": extractor,
"visualizer": visualizer,
"out_fname": args.output,
"entry_idx": 0,
}
return context
def create_argument_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(
description=DOC,
formatter_class=lambda prog: argparse.HelpFormatter(prog, max_help_position=120),
)
parser.set_defaults(func=lambda _: parser.print_help(sys.stdout))
subparsers = parser.add_subparsers(title="Actions")
for _, action in _ACTION_REGISTRY.items():
action.add_parser(subparsers)
return parser
def main():
parser = create_argument_parser()
args = parser.parse_args()
verbosity = args.verbosity if hasattr(args, "verbosity") else None
global logger
logger = setup_logger(name=LOGGER_NAME)
logger.setLevel(verbosity_to_level(verbosity))
args.func(args)
if __name__ == "__main__":
main()Editor is loading...