diff --git a/spineps/phase_instance.py b/spineps/phase_instance.py
index 4a1f383..f619c19 100755
--- a/spineps/phase_instance.py
+++ b/spineps/phase_instance.py
@@ -212,6 +212,7 @@ def get_corpus_coms(
 
     # Check against ivd order
     seg_sem = seg_nii.map_labels({Location.Endplate.value: Location.Vertebra_Disc.value}, verbose=False)
+    has_ivd: bool = Location.Vertebra_Disc.value in seg_sem.unique()
     subreg_cc, subreg_cc_n = seg_sem.get_segmentation_connected_components(labels=Location.Vertebra_Disc.value)
     subreg_cc = subreg_cc[Location.Vertebra_Disc.value]
     subreg_cc[subreg_cc > 0] += 100
@@ -242,7 +243,7 @@ def get_corpus_coms(
             continue
 
         nkey = stats_by_height_keys[nidx]
-        if nkey in stats_by_height and stats_by_height[nkey][1] == is_ivd:
+        if nkey in stats_by_height and stats_by_height[nkey][1] == is_ivd and has_ivd:
             neighbor = stats_by_height[nkey]
             neighborheight = neighbor[0]
             logger.print(
diff --git a/spineps/phase_pre.py b/spineps/phase_pre.py
index f92ac61..39214ac 100644
--- a/spineps/phase_pre.py
+++ b/spineps/phase_pre.py
@@ -14,6 +14,7 @@ def preprocess_input(
     mri_nii: NII,
     debug_data: dict,  # noqa: ARG001
     pad_size: int = 4,
+    proc_normalize_input: bool = True,
     proc_do_n4_bias_correction: bool = True,
     proc_crop_input: bool = True,
     verbose: bool = False,
@@ -24,8 +25,15 @@ def preprocess_input(
         # Crop Down
         try:
             # Enforce to range [0, 1500]
-            mri_nii.normalize_to_range_(min_value=0, max_value=9000, verbose=logger)
-            crop = mri_nii.compute_crop(dist=0) if proc_crop_input else (slice(None, None), slice(None, None), slice(None, None))
+            if proc_normalize_input:
+                mri_nii.normalize_to_range_(min_value=0, max_value=9000, verbose=logger)
+                crop = mri_nii.compute_crop(dist=0) if proc_crop_input else (slice(None, None), slice(None, None), slice(None, None))
+            else:
+                crop = (
+                    mri_nii.compute_crop(minimum=mri_nii.min(), dist=0)
+                    if proc_crop_input
+                    else (slice(None, None), slice(None, None), slice(None, None))
+                )
         except ValueError:
             logger.print("Image Nifty is empty, skip this", Log_Type.FAIL)
             return None, ErrCode.EMPTY
@@ -40,7 +48,8 @@ def preprocess_input(
             logger.print(f"N4 Bias field correction done in {perf_counter() - n4_start} sec", verbose=True)
 
         # Enforce to range [0, 1500]
-        cropped_nii.normalize_to_range_(min_value=0, max_value=1500, verbose=logger)
+        if proc_normalize_input:
+            cropped_nii.normalize_to_range_(min_value=0, max_value=1500, verbose=logger)
 
         # Uncrop again
         # uncropped_input[crop] = cropped_nii.get_array()
diff --git a/spineps/seg_enums.py b/spineps/seg_enums.py
index 24d89e1..a4ec94a 100755
--- a/spineps/seg_enums.py
+++ b/spineps/seg_enums.py
@@ -50,6 +50,7 @@ class Modality(Enum_Compare):
     CT = auto()
     SEG = auto()
     MPR = auto()
+    PD = auto()
 
     @classmethod
     def format_keys(cls, modalities: Self | list[Self]) -> list[str]:
diff --git a/spineps/seg_model.py b/spineps/seg_model.py
index 37d8666..07097e0 100755
--- a/spineps/seg_model.py
+++ b/spineps/seg_model.py
@@ -355,8 +355,8 @@ def run(
             targetc = targetc.to(torch.float32)
             logits = self.predictor.forward(targetc.to(self.device))
         #
-        except Exception:
-            # print("Channel-wise model failed, try legacy version")
+        except Exception as e:
+            print(f"Channel-wise model failed with {e}, try legacy version")
             do_backup = True
         #
         if do_backup:
diff --git a/spineps/seg_run.py b/spineps/seg_run.py
index d842c8d..07cb536 100755
--- a/spineps/seg_run.py
+++ b/spineps/seg_run.py
@@ -250,6 +250,7 @@ def process_img_nii(  # noqa: C901
     override_postpair: bool = False,
     override_ctd: bool = False,
     proc_pad_size: int = 4,
+    proc_normalize_input: bool = True,
     # Processings
     # Semantic
     proc_sem_crop_input: bool = True,
@@ -352,6 +353,9 @@ def process_img_nii(  # noqa: C901
     done_something = False
     debug_data_run: dict[str, NII] = {}
 
+    if Modality.CT in model_semantic.modalities():
+        proc_normalize_input = False  # Never normalize input if it is an CT
+
     compatible = check_input_model_compatibility(img_ref, model=model_semantic)
     if not compatible:
         if not ignore_compatibility_issues:
@@ -378,6 +382,7 @@ def process_img_nii(  # noqa: C901
                 pad_size=proc_pad_size,
                 debug_data=debug_data_run,
                 proc_crop_input=proc_sem_crop_input,
+                proc_normalize_input=proc_normalize_input,
                 proc_do_n4_bias_correction=proc_sem_n4_bias_correction,
                 verbose=verbose,
             )