def predict_file(self, csv_file, root_dir, savedir=None):
"""Create a dataset and predict entire annotation file
Csv file format is .csv file with the columns "image_path", "xmin","ymin","xmax","ymax" for the image name and bounding box position.
Image_path is the relative filename, not absolute path, which is in the root_dir directory. One bounding box per line.
Args:
csv_file: path to csv file
root_dir: directory of images. If none, uses "image_dir" in config
savedir: Optional. Directory to save image plots.
Returns:
df: pandas dataframe with bounding boxes, label and scores for each image in the csv file
"""
self.model.eval()
result = predict.predict_file(model=self.model,
csv_file=csv_file,
root_dir=root_dir,
savedir=savedir,
device=self.device,
iou_threshold=self.config["nms_thresh"])
#Set labels to character from numeric
result["label"] = result.label.apply(
lambda x: self.numeric_to_label_dict[x])
return result
def predict_file(self, csv_file, root_dir, savedir=None, color=None, thickness=1):
"""Create a dataset and predict entire annotation file
Csv file format is .csv file with the columns "image_path", "xmin","ymin","xmax","ymax" for the image name and bounding box position.
Image_path is the relative filename, not absolute path, which is in the root_dir directory. One bounding box per line.
Args:
csv_file: path to csv file
root_dir: directory of images. If none, uses "image_dir" in config
savedir: Optional. Directory to save image plots.
color: color of the bounding box as a tuple of BGR color, e.g. orange annotations is (0, 165, 255)
thickness: thickness of the rectangle border line in px
Returns:
df: pandas dataframe with bounding boxes, label and scores for each image in the csv file
"""
self.model = self.model.to(self.current_device)
self.model.eval()
self.model.score_thresh = self.config["score_thresh"]
result = predict.predict_file(model=self.model,
csv_file=csv_file,
root_dir=root_dir,
savedir=savedir,
device=self.current_device,
iou_threshold=self.config["nms_thresh"],
color=color,
thickness=thickness)
#Set labels to character from numeric
result["label"] = result.label.apply(lambda x: self.numeric_to_label_dict[x])
return result
def evaluate(self, csv_file, root_dir, iou_threshold=None, savedir=None):
"""Compute intersection-over-union and precision/recall for a given iou_threshold
Args:
csv_file: location of a csv file with columns "name","xmin","ymin","xmax","ymax","label", each box in a row
root_dir: location of files in the dataframe 'name' column.
iou_threshold: float [0,1] intersection-over-union union between annotation and prediction to be scored true positive
savedir: optional path dir to save evaluation images
Returns:
results: dict of ("results", "precision", "recall") for a given threshold
"""
# Load on GPU is available
if torch.cuda.is_available():
self.model = self.model.to("cuda")
self.model.eval()
self.model.score_thresh = self.config["score_thresh"]
predictions = predict.predict_file(
model=self.model,
csv_file=csv_file,
root_dir=root_dir,
savedir=savedir,
device=self.current_device,
iou_threshold=self.config["nms_thresh"])
ground_df = pd.read_csv(csv_file)
ground_df["label"] = ground_df.label.apply(
lambda x: self.label_dict[x])
#remove empty samples from ground truth
ground_df = ground_df[~((ground_df.xmin == 0) & (ground_df.xmax == 0))]
# if no arg for iou_threshold, set as config
if iou_threshold is None:
iou_threshold = self.config["validation"]["iou_threshold"]
results = evaluate_iou.evaluate(predictions=predictions,
ground_df=ground_df,
root_dir=root_dir,
iou_threshold=iou_threshold,
savedir=savedir)
#replace classes if not NUll, wrap in try catch if no predictions
if not results["results"].empty:
results["results"]["predicted_label"] = results["results"][
"predicted_label"].apply(lambda x: self.numeric_to_label_dict[
x] if not pd.isnull(x) else x)
results["results"]["true_label"] = results["results"][
"true_label"].apply(lambda x: self.numeric_to_label_dict[x])
results["predictions"] = predictions
return results
def log_images(self, pl_module):
boxes = predict.predict_file(model=pl_module.model,
csv_file=self.csv_file,
root_dir=self.root_dir,
savedir=self.savedir,
device=pl_module.device)
try:
saved_plots = glob.glob("{}/*.png".format(self.savedir))
for x in saved_plots:
pl_module.logger.experiment.log_image(x)
except Exception as e:
print(
"Could not find logger in lightning module, skipping upload, images were saved to {}, error was rasied {}"
.format(self.savedir, e))
def evaluate(self,
csv_file,
root_dir,
iou_threshold=None,
show_plot=False,
savedir=None):
"""Compute intersection-over-union and precision/recall for a given iou_threshold
Args:
df: a pandas-type dataframe (geopandas is fine) with columns "name","xmin","ymin","xmax","ymax","label", each box in a row
root_dir: location of files in the dataframe 'name' column.
iou_threshold: float [0,1] intersection-over-union union between annotation and prediction to be scored true positive
show_plot: open a blocking matplotlib window to show plot and annotations, useful for debugging.
savedir: optional path dir to save evaluation images
Returns:
results: dict of ("results", "precision", "recall") for a given threshold
"""
self.model.eval()
if not self.device.type == "cpu":
self.model = self.model.to(self.device)
predictions = predict.predict_file(
model=self.model,
csv_file=csv_file,
root_dir=root_dir,
savedir=savedir,
device=self.device,
iou_threshold=self.config["nms_thresh"])
predictions["label"] = predictions.label.apply(
lambda x: self.numeric_to_label_dict[x])
ground_df = pd.read_csv(csv_file)
# if no arg for iou_threshold, set as config
if iou_threshold is None:
iou_threshold = self.config["validation"]["iou_threshold"]
results = evaluate_iou.evaluate(predictions=predictions,
ground_df=ground_df,
root_dir=root_dir,
iou_threshold=iou_threshold,
show_plot=show_plot)
return results