Mirrors/blakeblackshear.frigate
1
0
Fork 0
mirror of https://github.com/blakeblackshear/frigate.git synced 2026-03-07 02:18:07 +01:00
Code Issues Projects Releases Wiki Activity

Generic classification card (#20379)

Browse Source 
* Refactor face card into generic classification card

* Update classification data card to use classification card

* Refactor state training grid to use classification card

* Refactor grouped face card into generic component

* Combine classification objects by event

* Fixup

* Cleanup

* Cleanup

* Do not fail if a single event is not found

* Save original frame

* Cleanup

* Undo
This commit is contained in:
Nicolas Mowen
2025-10-07 13:43:06 -06:00
committed by GitHub
parent 4bea69591b
commit 37afd5da6b
10 changed files with 705 additions and 452 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
frigate/data_processing/real_time/custom_classification.py
View File

@@ -142,7 +142,7 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
if frame.shape != (224, 224):
try:
frame = cv2.resize(frame, (224, 224))
resized_frame = cv2.resize(frame, (224, 224))
except Exception:
logger.warning("Failed to resize image for state classification")
return
@@ -151,13 +151,14 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
write_classification_attempt(
self.train_dir,
cv2.cvtColor(frame, cv2.COLOR_RGB2BGR),
"none-none",
now,
"unknown",
0.0,
)
return
input = np.expand_dims(frame, axis=0)
input = np.expand_dims(resized_frame, axis=0)
self.interpreter.set_tensor(self.tensor_input_details[0]["index"], input)
self.interpreter.invoke()
res: np.ndarray = self.interpreter.get_tensor(
@@ -171,6 +172,7 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
write_classification_attempt(
self.train_dir,
cv2.cvtColor(frame, cv2.COLOR_RGB2BGR),
"none-none",
now,
self.labelmap[best_id],
score,
@@ -284,7 +286,7 @@ class CustomObjectClassificationProcessor(RealTimeProcessorApi):
if crop.shape != (224, 224):
try:
crop = cv2.resize(crop, (224, 224))
resized_crop = cv2.resize(crop, (224, 224))
except Exception:
logger.warning("Failed to resize image for state classification")
return
@@ -293,13 +295,14 @@ class CustomObjectClassificationProcessor(RealTimeProcessorApi):
write_classification_attempt(
self.train_dir,
cv2.cvtColor(crop, cv2.COLOR_RGB2BGR),
obj_data["id"],
now,
"unknown",
0.0,
)
return
input = np.expand_dims(crop, axis=0)
input = np.expand_dims(resized_crop, axis=0)
self.interpreter.set_tensor(self.tensor_input_details[0]["index"], input)
self.interpreter.invoke()
res: np.ndarray = self.interpreter.get_tensor(
@@ -314,6 +317,7 @@ class CustomObjectClassificationProcessor(RealTimeProcessorApi):
write_classification_attempt(
self.train_dir,
cv2.cvtColor(crop, cv2.COLOR_RGB2BGR),
obj_data["id"],
now,
self.labelmap[best_id],
score,
@@ -372,6 +376,7 @@ class CustomObjectClassificationProcessor(RealTimeProcessorApi):
def write_classification_attempt(
folder: str,
frame: np.ndarray,
event_id: str,
timestamp: float,
label: str,
score: float,
@@ -379,7 +384,7 @@ def write_classification_attempt(
if "-" in label:
label = label.replace("-", "_")
file = os.path.join(folder, f"{timestamp}-{label}-{score}.webp")
file = os.path.join(folder, f"{event_id}-{timestamp}-{label}-{score}.webp")
os.makedirs(folder, exist_ok=True)
cv2.imwrite(file, frame)