update: api

common/api.py

@@ -39,7 +39,15 @@ class ImageMatchingAPI(torch.nn.Module):
                 "name": "xfeat",
                 "max_keypoints": 1024,
                 "keypoint_threshold": 0.015,
-            }
+            },
+            "preprocessing": {
+                "grayscale": False,
+                "resize_max": 1600,
+                "force_resize": True,
+                "width": 640,
+                "height": 480,
+                "dfactor": 8,
+            },
         },
         "ransac": {
             "enable": True,
@@ -75,12 +83,18 @@ class ImageMatchingAPI(torch.nn.Module):
         """
         super().__init__()
         self.device = device
-        self.conf = conf = {
-            **self.parse_match_config(self.default_conf),
-            **conf,
-        }
+        self.conf = self.parse_match_config(conf)
         self._updata_config(detect_threshold, max_keypoints, match_threshold)
         self._init_models()
+        if device == "cuda":
+            memory_allocated = torch.cuda.memory_allocated(device)
+            memory_reserved = torch.cuda.memory_reserved(device)
+            logger.info(
+                f"GPU memory allocated: {memory_allocated / 1024**2:.3f} MB"
+            )
+            logger.info(
+                f"GPU memory reserved: {memory_reserved / 1024**2:.3f} MB"
+            )
         self.pred = None
 
     def parse_match_config(self, conf):
@@ -123,7 +137,7 @@ class ImageMatchingAPI(torch.nn.Module):
 
     def _init_models(self):
         # initialize matcher
-        self.matcher = get_model(self.conf["matcher"])
+        self.matcher = get_model(self.match_conf)
         # initialize extractor
         if self.dense:
             self.extractor = None

test_app_cli.py

@@ -39,23 +39,57 @@ def test_one():
     img_path2 = ROOT / "datasets/sacre_coeur/mapping/17295357_9106075285.jpg"
     image0 = cv2.imread(str(img_path1))[:, :, ::-1]  # RGB
     image1 = cv2.imread(str(img_path2))[:, :, ::-1]  # RGB
-
+    # sparse
     conf = {
+        "dense": False,
         "matcher": {
-            "output": "matches-omniglue",
             "model": {
-                "name": "omniglue",
+                "name": "NN-mutual",
                 "match_threshold": 0.2,
-                "features": "null",
-            },
-            "preprocessing": {
-                "grayscale": False,
-                "resize_max": 1024,
-                "dfactor": 8,
-                "force_resize": False,
-            },
+            }
+        },
+        "feature": {
+            "model": {
+                "name": "xfeat",
+                "max_keypoints": 1024,
+                "keypoint_threshold": 0.015,
+            }
         },
+        "ransac": {
+            "enable": True,
+            "estimator": "poselib",
+            "geometry": "homography",
+            "method": "RANSAC",
+            "reproj_threshold": 3,
+            "confidence": 0.9999,
+            "max_iter": 10000,
+        },
+    }
+    api = ImageMatchingAPI(conf=conf, device=device)
+    api(image0, image1)
+    log_path = ROOT / "experiments" / "one"
+    log_path.mkdir(exist_ok=True, parents=True)
+    api.visualize(log_path=log_path)
+
+    # dense
+    conf = {
         "dense": True,
+        "matcher": {
+            "model": {
+                "name": "loftr",
+                "match_threshold": 0.2,
+            }
+        },
+        "feature": {},
+        "ransac": {
+            "enable": True,
+            "estimator": "poselib",
+            "geometry": "homography",
+            "method": "RANSAC",
+            "reproj_threshold": 3,
+            "confidence": 0.9999,
+            "max_iter": 10000,
+        },
     }
     api = ImageMatchingAPI(conf=conf, device=device)
     api(image0, image1)