Add margin check to multi-ref matching to reduce false positives

The multi-ref matching method was missing a margin check against other
logos, causing excessive false positives. This fix adds:

- margin parameter to find_best_match_multi_ref() that requires the
  best logo's score to exceed the second-best by a minimum margin
- Test script now passes --margin to both matching methods
- Updated documentation to reflect margin applies to both methods

Also adds run_comparison_tests.sh to run all three matching methods
and compare results.

test_logo_detection.py

@@ -231,7 +231,7 @@ def main():
         "--margin",
         type=float,
         default=0.05,
-        help="Required margin between best and second-best match for 'margin' method (default: 0.05)",
+        help="Required margin between best and second-best match (applies to both methods) (default: 0.05)",
     )
     parser.add_argument(
         "--matching-method",
@@ -453,6 +453,7 @@ def main():
                     similarity_threshold=args.threshold,
                     min_matching_refs=args.min_matching_refs,
                     use_mean_similarity=not args.use_max_similarity,
+                    margin=args.margin,
                 )
                 if match_result:
                     label, similarity, num_matching = match_result
@@ -511,9 +512,8 @@ def main():
     print(f"  CLIP similarity threshold: {args.threshold}")
     print(f"  DETR confidence threshold: {args.detr_threshold}")
     print(f"  Matching method:           {args.matching_method}")
-    if args.matching_method == "margin":
-        print(f"  Matching margin:           {args.margin}")
-    else:  # multi-ref
+    print(f"  Matching margin:           {args.margin}")
+    if args.matching_method == "multi-ref":
         print(f"  Min matching refs:         {args.min_matching_refs}")
         print(f"  Similarity aggregation:    {'max' if args.use_max_similarity else 'mean'}")
     if args.seed is not None: