diff --git a/arcos4py/tools/_detect_events.py b/arcos4py/tools/_detect_events.py
index 14f25de..7f03504 100644
--- a/arcos4py/tools/_detect_events.py
+++ b/arcos4py/tools/_detect_events.py
@@ -354,6 +354,41 @@ def reset(self):
         self.coordinates = []
         self.prev_cluster_ids = []
 
+    def overwrite_cluster_ids(self, seed_coordinates: np.ndarray, seed_cluster_ids: np.ndarray):
+        """Adds seed clusters to the memory.
+
+        Arguments:
+            seed_coordinates (np.ndarray): Array of shape (N, D) of seed coordinates.
+            seed_cluster_ids (np.ndarray): Array of shape (N,) of seed cluster IDs.
+        """
+        # convert to int
+        print("Overwriting cluster IDs in memory")
+        seed_cluster_ids = seed_cluster_ids.astype(int)
+        seed_coordinates = seed_coordinates.astype(int)
+        if not self.coordinates:
+            valid = seed_cluster_ids > 0
+            self.coordinates.append(seed_coordinates[valid])
+            self.prev_cluster_ids.append(seed_cluster_ids[valid])
+            self.max_prev_cluster_id = np.max(seed_cluster_ids[valid])
+        else:
+            coords = self.coordinates[-1]
+            ids = self.prev_cluster_ids[-1]
+            for sc, sci in zip(seed_coordinates, seed_cluster_ids):
+                if sci <= 0:
+                    continue
+                if coords.size > 0:
+                    matches = np.all(coords == sc, axis=1)
+                    idx = np.nonzero(matches)[0]
+                    if idx.size > 0:
+                        ids[idx[0]] = sci
+                        self.max_prev_cluster_id = max(self.max_prev_cluster_id, sci)
+                        continue
+                coords = np.vstack([coords, sc[np.newaxis, :]]) if coords.size > 0 else sc[np.newaxis, :]
+                ids = np.concatenate([ids, [sci]]) if ids.size > 0 else np.array([sci], dtype=int)
+                self.max_prev_cluster_id = max(self.max_prev_cluster_id, sci)
+            self.coordinates[-1] = coords
+            self.prev_cluster_ids[-1] = ids
+
     @property
     def all_coordinates(self):
         """Returns all coordinates in memory as one array."""
@@ -1183,8 +1218,6 @@ def _update_tree(self, coords):
     def _get_next_id(self) -> int:
         """Generate a new unique ID."""
         self._memory.max_prev_cluster_id += 1
-        if self._memory.max_prev_cluster_id == 92:
-            pass
         return self._memory.max_prev_cluster_id
 
     def _apply_remove_small_clusters(self, linked_cluster_ids, original_cluster_ids):
@@ -1397,6 +1430,23 @@ def _apply_stable_merges_splits(self, linked_cluster_ids, original_cluster_ids):
 
         return final_cluster_ids
 
+    def add_seed_cluster_to_memory(self, cluster_id: np.ndarray, coordinates: np.ndarray):
+        """Adds a seed cluster to memory or overwrites an existing cluster.
+
+        Arguments:
+            cluster_id (np.ndarray): The ID of the cluster.
+            coordinates (np.ndarray): The coordinates of the cluster.
+        """
+        self._memory.overwrite_cluster_ids(cluster_id, coordinates)
+        memory_coordinates = self._memory.coordinates
+        if len(memory_coordinates) > 1:
+            _memory_coordinates = np.concatenate(memory_coordinates)
+        elif len(memory_coordinates) == 1:
+            _memory_coordinates = memory_coordinates[0]
+        else:
+            raise ValueError("Memory coordinates are empty")
+        self._update_tree(_memory_coordinates)
+
 
 class BaseTracker(ABC):
     """Abstract base class for tracker classes."""
@@ -1678,6 +1728,29 @@ def _coordinates_to_image(self, x, position_data, tracked_events):
 
         return out_img
 
+    def add_seed_image(self, seed_image: np.ndarray) -> None:
+        """Adds a seed image to the tracker. This is used to guide the tracking process\
+        with a specific starting label image.
+
+        Arguments:
+            seed_image (np.ndarray): The seed image to add. Should be a 2D array of integers.
+        """
+        if seed_image.ndim != 2:
+            raise ValueError("Seed image must be 2D.")
+        if not np.issubdtype(seed_image.dtype, np.integer):
+            raise ValueError("Seed image must be a label image (integer type).")
+        if self.linker.frame_counter != -1:
+            warnings.warn(
+                "Seed image is intended to be used before any tracking is done. If\
+                used after first frame coordinates in memory will be overwriten if present."
+            )
+
+        seed_image = downscale_image(seed_image, self._downsample)
+
+        coords = self._image_to_coordinates(seed_image)
+
+        self.linker.add_seed_cluster_to_memory(coords[0], coords[1])
+
     def track_iteration(self, x: np.ndarray) -> np.ndarray:
         """Tracks events in a single frame. Returns the tracked labels.