Merge pull request #61 from ethanweber/tancik/colormap_logs

Add colormaps to logs

.vscode/settings.json

@@ -10,6 +10,9 @@
     },
     "editor.formatOnSave": true,
     "python.formatting.provider": "black",
+    "python.formatting.blackArgs": [
+        "--line-length=120"
+    ],
     "python.linting.pylintEnabled": true,
     "python.linting.flake8Enabled": false,
     "python.linting.enabled": true,

environment/requirements.txt

@@ -5,6 +5,7 @@ imageio==2.16.1
 ipywidgets>=7.6
 jupyterlab
 kaleido
+matplotlib
 mediapy
 plotly==5.7.0
 pylint==2.13.4

mattport/nerf/graph/vanilla_nerf.py

@@ -20,6 +20,7 @@ from mattport.nerf.renderers import AccumulationRenderer, DisparityRenderer, RGB
 from mattport.nerf.sampler import PDFSampler, UniformSampler
 from mattport.structures import colors
 from mattport.structures.rays import RaySamples
+from mattport.utils import visualization
 
 
 class NeRFField(nn.Module):
@@ -162,12 +163,16 @@ class NeRFGraph(Graph):
         fine_renderer_accumulation = self.renderer_accumulation(fine_weights)  # RendererOutputs
         fine_renderer_disparity = self.renderer_disparity(fine_weights, pdf_ray_samples.ts)
 
+        # TODO refactor this into "vis" section. Doesn't need to be run during training.
+        coarse_renderer_accumulation = visualization.apply_colormap(coarse_renderer_accumulation.accumulation)
+        fine_renderer_accumulation = visualization.apply_colormap(fine_renderer_accumulation.accumulation)
+
         # outputs:
         outputs = {
             "rgb_coarse": coarse_renderer_outputs.rgb,
             "rgb_fine": fine_renderer_outputs.rgb,
-            "accumulation_coarse": coarse_renderer_accumulation.accumulation,
-            "accumulation_fine": fine_renderer_accumulation.accumulation,
+            "accumulation_coarse": coarse_renderer_accumulation,
+            "accumulation_fine": fine_renderer_accumulation,
             "disparity_coarse": coarse_renderer_disparity.disparity,
             "disparity_fine": fine_renderer_disparity.disparity,
         }

mattport/nerf/trainer.py

@@ -260,8 +260,8 @@ class Trainer:
                 disparity_fine.append(graph_outputs["disparity_fine"])
             rgb_coarse = torch.cat(rgb_coarse).view(image_height, image_width, 3).detach().cpu()
             rgb_fine = torch.cat(rgb_fine).view(image_height, image_width, 3).detach().cpu()
-            accumulation_coarse = torch.cat(accumulation_coarse).view(image_height, image_width, 1).detach().cpu()
-            accumulation_fine = torch.cat(accumulation_fine).view(image_height, image_width, 1).detach().cpu()
+            accumulation_coarse = torch.cat(accumulation_coarse).view(image_height, image_width, 3).detach().cpu()
+            accumulation_fine = torch.cat(accumulation_fine).view(image_height, image_width, 3).detach().cpu()
             disparity_coarse = torch.cat(disparity_coarse).view(image_height, image_width, 1).detach().cpu()
             disparity_fine = torch.cat(disparity_fine).view(image_height, image_width, 1).detach().cpu()
 

mattport/utils/visualization.py

@@ -0,0 +1,24 @@
+""" Helper functions for visualizing outputs """
+
+import torch
+from matplotlib import cm
+from torchtyping import TensorType
+
+
+def apply_colormap(image: TensorType[..., 1], cmap="viridis") -> TensorType[..., 3]:
+    """Convert single channel to a color image.
+
+    Args:
+        image (TensorType[..., 1]): Single channel image.
+        cmap (str, optional): Colormap for image. Defaults to 'turbo'.
+
+    Returns:
+        TensorType[..., 3]: Colored image
+    """
+
+    colormap = cm.get_cmap(cmap)
+    colormap = torch.tensor(colormap.colors).to(image.device)
+
+    image = (image * 255).long()
+
+    return colormap[image]

mattport/utils/writer.py

@@ -89,13 +89,13 @@ class TensorboardWriter(Writer):
             self.tb_writer = SummaryWriter(log_dir=self.save_dir)
 
     def write_image(
-        self, name: str, x: TensorType["H", "W", 3], step: int, group: str = None, prefix: str = None
+        self, name: str, x: TensorType["H", "W", "C"], step: int, group: str = None, prefix: str = None
     ) -> None:
         """_summary_
 
         Args:
             name (str): data identifier
-            x (TensorType["H", "W", 3]): rendered image to write
+            x (TensorType["H", "W", "C"]): rendered image to write
         """
         x = to8b(x)
         tensorboard_name = get_tensorboard_name(name, group=group, prefix=prefix)