dave/py-media-grader
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor and streamline template matching logic in VideoEditor

Browse Source 
This commit removes unused methods and simplifies the template matching process by eliminating multi-scale matching in favor of a single-scale approach. The changes enhance performance and clarity in the tracking logic, ensuring a more efficient template matching experience. Debug messages have been updated to reflect the current state of the template matching process.
This commit is contained in:
PhatPhuckDave
2025-09-26 18:30:19 +02:00
parent 44ed4220b9
commit 01340a0a81
1 changed files with 12 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files

114
croppa/main.py
View File

@@ -4,7 +4,7 @@ import cv2
import argparse
import numpy as np
from pathlib import Path
from typing import List, Optional, Tuple, Dict, Any
from typing import List, Dict, Any
import time
import re
import threading
@@ -200,24 +200,6 @@ class FeatureTracker:
def get_tracking_position(self, frame_number: int) -> Optional[Tuple[float, float]]:
"""Get the average tracking position for a frame"""
if frame_number not in self.features:
return None
if not self.features[frame_number]['positions']:
return None
positions = self.features[frame_number]['positions']
if not positions:
return None
# Calculate average position
avg_x = sum(pos[0] for pos in positions) / len(positions)
avg_y = sum(pos[1] for pos in positions) / len(positions)
return (avg_x, avg_y)
def clear_features(self):
@@ -2308,17 +2290,6 @@ class VideoEditor:
return (interp_x, interp_y)
def set_tracking_template(self, frame, region):
"""Set a template region for tracking (much better than optical flow)"""
try:
x, y, w, h = region
self.tracking_template = frame[y:y+h, x:x+w].copy()
self.template_region = region
print(f"DEBUG: Set tracking template with region {region}")
return True
except Exception as e:
print(f"Error setting tracking template: {e}")
return False
def track_template(self, frame):
"""Track the template in the current frame"""
@@ -2329,48 +2300,19 @@ class VideoEditor:
# Apply image preprocessing for better template matching
gray_frame, gray_template = self._improve_template_matching(frame, self.tracking_template)
# Multi-scale template matching for better tracking (if enabled)
if False: # Multi-scale template matching removed
scales = [0.8, 0.9, 1.0, 1.1, 1.2] # Different scales to try
# Single-scale template matching (faster)
result = cv2.matchTemplate(gray_frame, gray_template, cv2.TM_CCOEFF_NORMED)
_, max_val, _, max_loc = cv2.minMaxLoc(result)
if max_val > 0.6: # Higher threshold for single-scale
template_h, template_w = gray_template.shape
center_x = max_loc[0] + template_w // 2
center_y = max_loc[1] + template_h // 2
best_match = (center_x, center_y, max_val)
best_confidence = max_val
else:
best_match = None
best_confidence = 0.0
for scale in scales:
# Resize template
template_h, template_w = gray_template.shape
new_w = int(template_w * scale)
new_h = int(template_h * scale)
if new_w <= 0 or new_h <= 0 or new_w > gray_frame.shape[1] or new_h > gray_frame.shape[0]:
continue
scaled_template = cv2.resize(gray_template, (new_w, new_h))
# Perform template matching
result = cv2.matchTemplate(gray_frame, scaled_template, cv2.TM_CCOEFF_NORMED)
_, max_val, _, max_loc = cv2.minMaxLoc(result)
# Check if this is the best match so far
if max_val > best_confidence:
best_confidence = max_val
# Get center of template
center_x = max_loc[0] + new_w // 2
center_y = max_loc[1] + new_h // 2
best_match = (center_x, center_y, max_val)
else:
# Single-scale template matching (faster)
result = cv2.matchTemplate(gray_frame, gray_template, cv2.TM_CCOEFF_NORMED)
_, max_val, _, max_loc = cv2.minMaxLoc(result)
if max_val > 0.6: # Higher threshold for single-scale
template_h, template_w = gray_template.shape
center_x = max_loc[0] + template_w // 2
center_y = max_loc[1] + template_h // 2
best_match = (center_x, center_y, max_val)
best_confidence = max_val
else:
best_match = None
best_confidence = 0.0
# Adaptive thresholding based on recent match history
if len(self.template_match_history) > 0:
@@ -2394,36 +2336,6 @@ class VideoEditor:
print(f"Error in template tracking: {e}")
return None
def _recreate_template_from_region(self, frame):
"""Recreate template from saved region coordinates"""
try:
if self.template_region is None:
return False
x, y, w, h = self.template_region
print(f"DEBUG: Recreating template from region ({x}, {y}, {w}, {h})")
# Ensure region is within frame bounds
if (x >= 0 and y >= 0 and
x + w <= frame.shape[1] and
y + h <= frame.shape[0]):
# Extract template from frame
template = frame[y:y+h, x:x+w]
if template.size > 0:
self.tracking_template = template.copy()
print(f"DEBUG: Template recreated with size {template.shape}")
return True
else:
print("DEBUG: Template region too small")
return False
else:
print("DEBUG: Template region outside frame bounds")
return False
except Exception as e:
print(f"Error recreating template: {e}")
return False
def _improve_template_matching(self, frame, template):
"""Apply image preprocessing to improve template matching"""
@@ -3253,8 +3165,6 @@ class VideoEditor:
if self.template_matching_enabled:
mode = "Full Frame" if self.template_matching_full_frame else "Cropped"
template_text = f" | Template: {mode}"
if False: # Multi-scale template matching removed
template_text += " (MULTI-SCALE)"
autorepeat_text = (
f" | Loop: ON" if self.looping_between_markers else ""
)