diff --git a/lightglue/__init__.py b/lightglue/__init__.py
index 42719c9..95c3654 100644
--- a/lightglue/__init__.py
+++ b/lightglue/__init__.py
@@ -3,4 +3,5 @@
 from .lightglue import LightGlue  # noqa
 from .sift import SIFT  # noqa
 from .superpoint import SuperPoint  # noqa
+from .dog_hardnet import DoGHardNet  # noqa
 from .utils import match_pair  # noqa
diff --git a/lightglue/dog_hardnet.py b/lightglue/dog_hardnet.py
index e53e97a..8c7e465 100644
--- a/lightglue/dog_hardnet.py
+++ b/lightglue/dog_hardnet.py
@@ -1,41 +1,19 @@
-import warnings
-
-import cv2
-import numpy as np
 import torch
 from kornia.color import rgb_to_grayscale
 from kornia.feature import HardNet, LAFDescriptor, laf_from_center_scale_ori
-from packaging import version
 
-from .utils import Extractor
 from .sift import SIFT
 
-try:
-    import pycolmap
-except ImportError:
-    pycolmap = None
-
 
 class DoGHardNet(SIFT):
-    default_conf = {
-        "nms_radius": 0,  # None to disable filtering entirely.
-        "max_num_keypoints": 2048,
-        "backend": "opencv",  # in {opencv, pycolmap, pycolmap_cpu, pycolmap_cuda}
-        "detection_threshold": -1,  # from COLMAP
-        "edge_threshold": -1,
-        "first_octave": -1,  # only used by pycolmap, the default of COLMAP
-        "num_octaves": 4,
-        "force_num_keypoints": True,
-    }
-
     required_data_keys = ["image"]
 
-    def _init(self, conf):
-        super()._init(conf)
+    def __init__(self, **conf):
+        super().__init__(**conf)
         self.laf_desc = LAFDescriptor(HardNet(True)).eval()
 
 
-    def _forward(self, data: dict) -> dict:
+    def forward(self, data: dict) -> dict:
         image = data["image"]
         if image.shape[1] == 3:
             image = rgb_to_grayscale(image)
@@ -43,7 +21,10 @@ def _forward(self, data: dict) -> dict:
         self.laf_desc = self.laf_desc.to(device)
         self.laf_desc.descriptor = self.laf_desc.descriptor.eval()
         pred = []
-        im_size = data.get("image_size").long()
+        if "image_size" in data.keys():
+            im_size = data.get("image_size").long()
+        else:
+            im_size = None
         for k in range(len(image)):
             img = image[k]
             if im_size is not None:
@@ -58,8 +39,4 @@ def _forward(self, data: dict) -> dict:
             p["descriptors"] = self.laf_desc(img[None], lafs).reshape(-1, 128)
             pred.append(p)
         pred = {k: torch.stack([p[k] for p in pred], 0).to(device) for k in pred[0]}
-        return pred
-
-
-
-        
\ No newline at end of file
+        return pred
\ No newline at end of file