diff --git a/cvat/apps/quality_control/views.py b/cvat/apps/quality_control/views.py
index 2d86ef586847..2190ffce9ef6 100644
--- a/cvat/apps/quality_control/views.py
+++ b/cvat/apps/quality_control/views.py
@@ -446,10 +446,18 @@ def immediate_reports(self, request, *args, **kwargs):
                 # handle label matching task
                 else:
                     penalty_factor = 0.1
-                    def calculate_score(gt_samples, ds_samples):
-                        # Group annotations by label
+                    def calculate_score(gt_samples, ds_samples, start_time=0):
+                        gt_samples_adjusted = []
+                        for ann in gt_samples:
+                            adjusted_ann = ann.copy()
+                            adjusted_points = adjusted_ann["points"].copy()
+                            adjusted_points[0] = round(adjusted_points[0] - start_time, 10)  # Round to 10 decimal places
+                            adjusted_points[3] = round(adjusted_points[3] - start_time, 10)
+                            adjusted_ann["points"] = adjusted_points
+                            gt_samples_adjusted.append(adjusted_ann)
+
                         gt_by_label = {}
-                        for idx, ann in enumerate(gt_samples):
+                        for idx, ann in enumerate(gt_samples_adjusted):
                             label = ann.get("label_id", "")
                             if label not in gt_by_label:
                                 gt_by_label[label] = []
@@ -466,27 +474,22 @@ def calculate_score(gt_samples, ds_samples):
                         total_gt_count = 0
                         unused_predictions = []
 
-                        # Process each label separately
                         for label in set(list(gt_by_label.keys()) + list(ds_by_label.keys())):
                             gt_list = gt_by_label.get(label, [])
                             ds_list = ds_by_label.get(label, [])
                             total_gt_count += len(gt_list)
 
-                            # Track best coverage for each GT
                             gt_coverage = [0.0] * len(gt_list)
                             used_predictions = [False] * len(ds_list)
 
-                            # Check each GT against all predictions
                             for gt_idx, (orig_gt_idx, gt_ann) in enumerate(gt_list):
                                 gt_start, gt_end = gt_ann["points"][0], gt_ann["points"][3]
                                 gt_length = gt_end - gt_start
 
-                                # Find all overlapping predictions
                                 overlapping_predictions = []
                                 for ds_idx, ds_ann in enumerate(ds_list):
                                     ds_start, ds_end = ds_ann["points"][0], ds_ann["points"][3]
 
-                                    # Calculate overlap
                                     overlap_start = max(gt_start, ds_start)
                                     overlap_end = min(gt_end, ds_end)
                                     overlap = max(0, overlap_end - overlap_start)
@@ -495,7 +498,6 @@ def calculate_score(gt_samples, ds_samples):
                                         overlapping_predictions.append((ds_idx, overlap))
                                         used_predictions[ds_idx] = True
 
-                                # Merge overlaps to calculate total coverage
                                 overlapping_predictions.sort(key=lambda x: x[1], reverse=True)
                                 covered_intervals = []
 
@@ -520,7 +522,7 @@ def calculate_score(gt_samples, ds_samples):
 
                                 # Calculate total coverage
                                 total_covered = sum(end - start for start, end in covered_intervals)
-                                coverage_ratio = min(1.0, total_covered / gt_length)
+                                coverage_ratio = min(1.0, total_covered / gt_length) if gt_length > 0 else 0
                                 gt_coverage[gt_idx] = coverage_ratio
 
                             # Add unused predictions for this label to global list
@@ -547,7 +549,7 @@ def calculate_score(gt_samples, ds_samples):
 
                         return final_score
 
-                    score = calculate_score(gt_samples_filtered, ds_samples_filtered)
+                    score = calculate_score(gt_samples_filtered, ds_samples_filtered, start_time)
 
                 response_data = {
                     "score": score,