autoforge 1.3.1__tar.gz → 1.3.2__tar.gz

{autoforge-1.3.1 → autoforge-1.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: AutoForge
-Version: 1.3.1
+Version: 1.3.2
 Summary: AutoForge is a Python tool for generating 3D printed layered models from an input image.
 Author-email: Hendric Voss <hvoss@techfak.uni-bielefeld.de>
 License: CC BY-NC-SA 4.0
@@ -162,7 +162,6 @@ Although I would love to be it fully color compatible with hueforge, I don't thi
 AutoForge © 2025 by Hendric Voss is licensed under [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/).
 The software is provided as-is and comes with no warranty or guarantee of support.
 
-The 
 
 ## Acknowledgements
 

{autoforge-1.3.1 → autoforge-1.3.2}/README.md

@@ -142,7 +142,6 @@ Although I would love to be it fully color compatible with hueforge, I don't thi
 AutoForge © 2025 by Hendric Voss is licensed under [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/).
 The software is provided as-is and comes with no warranty or guarantee of support.
 
-The 
 
 ## Acknowledgements
 

{autoforge-1.3.1 → autoforge-1.3.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "AutoForge"
-version="1.3.1"
+version="1.3.2"
 description = "AutoForge is a Python tool for generating 3D printed layered models from an input image."
 authors = [{ name = "Hendric Voss", email = "hvoss@techfak.uni-bielefeld.de" }]
 readme = "README.md"

{autoforge-1.3.1 → autoforge-1.3.2}/src/AutoForge.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: AutoForge
-Version: 1.3.1
+Version: 1.3.2
 Summary: AutoForge is a Python tool for generating 3D printed layered models from an input image.
 Author-email: Hendric Voss <hvoss@techfak.uni-bielefeld.de>
 License: CC BY-NC-SA 4.0
@@ -162,7 +162,6 @@ Although I would love to be it fully color compatible with hueforge, I don't thi
 AutoForge © 2025 by Hendric Voss is licensed under [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/).
 The software is provided as-is and comes with no warranty or guarantee of support.
 
-The 
 
 ## Acknowledgements
 

{autoforge-1.3.1 → autoforge-1.3.2}/src/AutoForge.egg-info/SOURCES.txt

@@ -10,7 +10,6 @@ src/autoforge/__init__.py
 src/autoforge/__main__.py
 src/autoforge/auto_forge.py
 src/autoforge/auto_forge_gradio.py
-src/autoforge/color_fit_bayesian.py
 src/autoforge/filamentcolors_library.py
 src/autoforge/old_jax_files.py
 src/autoforge/Helper/FilamentHelper.py
@@ -22,5 +21,7 @@ src/autoforge/Helper/__init__.py
 src/autoforge/Loss/LossFunctions.py
 src/autoforge/Loss/PerceptionLoss.py
 src/autoforge/Loss/__init__.py
+src/autoforge/Misc/__init__.py
+src/autoforge/Misc/color_fit_all.py
 src/autoforge/Modules/Optimizer.py
 src/autoforge/Modules/__init__.py

autoforge-1.3.2/src/autoforge/Helper/ImageHelper.py

@@ -0,0 +1,22 @@
+import cv2
+
+
+def resize_image(img, max_size):
+    h_img, w_img, _ = img.shape
+
+    # Compute the scaling factor based on the larger dimension
+    if w_img >= h_img:
+        scale = max_size / w_img
+    else:
+        scale = max_size / h_img
+
+    # Compute new dimensions and round them
+    new_w = int(round(w_img * scale))
+    new_h = int(round(h_img * scale))
+
+    # Resize the image with an appropriate interpolation method
+    img_out = cv2.resize(img, (new_w, new_h), interpolation=cv2.INTER_CUBIC)
+    # INTER_CUBIC is actually not a good choice for shrinking images. Except in our case.
+    # As we shrink the image, we lose information and the image becomes blurry with INTER_AREA. This is detrimental for the solver. If you want to try it out be my guest.
+    # INTER_AREA actually destroys a lot of the colors of our solver.
+    return img_out

{autoforge-1.3.1 → autoforge-1.3.2}/src/autoforge/Helper/OptimizerHelper.py

@@ -1,4 +1,5 @@
 import random
+from itertools import permutations
 
 import numpy as np
 import torch
@@ -7,6 +8,7 @@ from skimage.color import rgb2lab
 from sklearn.cluster import KMeans
 from sklearn.exceptions import ConvergenceWarning
 from sklearn.utils._testing import ignore_warnings
+from tqdm import tqdm
 from transformers import pipeline
 from PIL import Image
 from sklearn.metrics import silhouette_score
@@ -110,9 +112,10 @@ def composite_image(
     comp = torch.zeros(H, W, 3, dtype=torch.float32, device=pixel_height.device)
     remaining = torch.ones(H, W, dtype=torch.float32, device=pixel_height.device)
 
-    A = 0.1215
-    k = 61.6970
-    b = 0.4773
+    o = -1.1021105e-02
+    A = 1.0786635e00
+    k = 3.0878708e01
+    b = -5.0812101e00
 
     for i in range(max_layers):
         layer_idx = max_layers - 1 - i
@@ -146,8 +149,11 @@ def composite_image(
         color_i = torch.matmul(p_i, material_colors)
         TD_i = torch.matmul(p_i, material_TDs) * 0.1
         TD_i = torch.clamp(TD_i, 1e-8, 1e8)
-        opac = A * torch.log1p(k * (eff_thick / TD_i)) + b * (eff_thick / TD_i)
+
+        opac = o + (A * torch.log1p(k * (eff_thick / TD_i)) + b * (eff_thick / TD_i))
         opac = torch.clamp(opac, 0.0, 1.0)
+        # Setting opac to 1.0 for the layers where the thickness is less than 2*h to avoid single layer artifacts
+        opac = torch.where(TD_i <= h * 2, torch.tensor(1.0), opac)
 
         comp = comp + ((remaining * opac).unsqueeze(-1) * color_i)
         remaining = remaining * (1 - opac)
@@ -610,7 +616,7 @@ def init_height_map(target, max_layers, h, eps=1e-6, random_seed=None):
 
     # --- Step 2: Second clustering of centroids into bands ---
     num_bands = choose_optimal_num_bands(
-        centroids, min_bands=5, max_bands=max_layers // 2, random_seed=random_seed
+        centroids, min_bands=5, max_bands=10, random_seed=random_seed
     )
     band_kmeans = KMeans(n_clusters=num_bands, random_state=random_seed).fit(centroids)
     band_labels = band_kmeans.labels_
@@ -643,7 +649,39 @@ def init_height_map(target, max_layers, h, eps=1e-6, random_seed=None):
     L_values = [lab[0] for lab in band_reps]
     start_band = np.argmin(L_values)  # darkest band index
     end_band = np.argmax(L_values)  # brightest band index
-    best_order = tsp_simulated_annealing(band_reps, start_band, end_band)
+
+    # --- Step 5: Find the best ordering for the middle bands ---
+    # We want to order the bands so that the total perceptual difference (Euclidean distance in Lab)
+    # between consecutive bands is minimized, while forcing the darkest band first and brightest band last.
+    all_indices = list(range(len(bands)))
+    middle_indices = [i for i in all_indices if i not in (start_band, end_band)]
+
+    min_total_distance = np.inf
+    best_order = None
+    total = len(middle_indices) * len(middle_indices)
+    # Try all permutations of the middle bands
+    ie = 0
+    tbar = tqdm(
+        permutations(middle_indices),
+        total=total,
+        desc="Finding best ordering for color bands:",
+    )
+    for perm in tbar:
+        candidate = [start_band] + list(perm) + [end_band]
+        total_distance = 0
+        for i in range(len(candidate) - 1):
+            total_distance += np.linalg.norm(
+                band_reps[candidate[i]] - band_reps[candidate[i + 1]]
+            )
+        if total_distance < min_total_distance:
+            min_total_distance = total_distance
+            best_order = candidate
+            tbar.set_description(
+                f"Finding best ordering for color bands: Total distance = {min_total_distance:.2f}"
+            )
+        ie += 1
+        if ie > 500000:
+            break
 
     new_order = []
     for band_idx in best_order:
@@ -656,7 +694,7 @@ def init_height_map(target, max_layers, h, eps=1e-6, random_seed=None):
 
     new_labels = new_labels.astype(np.float32) / new_labels.max()
 
-    normalized_lum = np.array(new_labels, dtype=np.float64)
+    normalized_lum = np.array(new_labels, dtype=np.float32)
     # convert out to inverse sigmoid logit function
     pixel_height_logits = np.log((normalized_lum + eps) / (1 - normalized_lum + eps))