Source code for nerfstudio.fields.vanilla_nerf_field

# Copyright 2022 the Regents of the University of California, Nerfstudio Team and contributors. All rights reserved.
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# See the License for the specific language governing permissions and
# limitations under the License.

"""Classic NeRF field"""

from typing import Dict, Optional, Tuple, Type

import torch
from torch import Tensor, nn

from nerfstudio.cameras.rays import RaySamples
from nerfstudio.field_components.encodings import Encoding, Identity
from nerfstudio.field_components.field_heads import DensityFieldHead, FieldHead, FieldHeadNames, RGBFieldHead
from nerfstudio.field_components.mlp import MLP
from nerfstudio.field_components.spatial_distortions import SpatialDistortion
from nerfstudio.fields.base_field import Field

[docs]class NeRFField(Field): """NeRF Field Args: position_encoding: Position encoder. direction_encoding: Direction encoder. base_mlp_num_layers: Number of layers for base MLP. base_mlp_layer_width: Width of base MLP layers. head_mlp_num_layers: Number of layer for output head MLP. head_mlp_layer_width: Width of output head MLP layers. skip_connections: Where to add skip connection in base MLP. use_integrated_encoding: Used integrated samples as encoding input. spatial_distortion: Spatial distortion. """ def __init__( self, position_encoding: Encoding = Identity(in_dim=3), direction_encoding: Encoding = Identity(in_dim=3), base_mlp_num_layers: int = 8, base_mlp_layer_width: int = 256, head_mlp_num_layers: int = 2, head_mlp_layer_width: int = 128, skip_connections: Tuple[int] = (4,), field_heads: Optional[Tuple[Type[FieldHead]]] = (RGBFieldHead,), use_integrated_encoding: bool = False, spatial_distortion: Optional[SpatialDistortion] = None, ) -> None: super().__init__() self.position_encoding = position_encoding self.direction_encoding = direction_encoding self.use_integrated_encoding = use_integrated_encoding self.spatial_distortion = spatial_distortion self.mlp_base = MLP( in_dim=self.position_encoding.get_out_dim(), num_layers=base_mlp_num_layers, layer_width=base_mlp_layer_width, skip_connections=skip_connections, out_activation=nn.ReLU(), ) self.field_output_density = DensityFieldHead(in_dim=self.mlp_base.get_out_dim()) if field_heads: self.mlp_head = MLP( in_dim=self.mlp_base.get_out_dim() + self.direction_encoding.get_out_dim(), num_layers=head_mlp_num_layers, layer_width=head_mlp_layer_width, out_activation=nn.ReLU(), ) self.field_heads = nn.ModuleList([field_head() for field_head in field_heads] if field_heads else []) # type: ignore for field_head in self.field_heads: field_head.set_in_dim(self.mlp_head.get_out_dim()) # type: ignore
[docs] def get_density(self, ray_samples: RaySamples) -> Tuple[Tensor, Tensor]: if self.use_integrated_encoding: gaussian_samples = ray_samples.frustums.get_gaussian_blob() if self.spatial_distortion is not None: gaussian_samples = self.spatial_distortion(gaussian_samples) encoded_xyz = self.position_encoding(gaussian_samples.mean, covs=gaussian_samples.cov) else: positions = ray_samples.frustums.get_positions() if self.spatial_distortion is not None: positions = self.spatial_distortion(positions) encoded_xyz = self.position_encoding(positions) base_mlp_out = self.mlp_base(encoded_xyz) density = self.field_output_density(base_mlp_out) return density, base_mlp_out
[docs] def get_outputs( self, ray_samples: RaySamples, density_embedding: Optional[Tensor] = None ) -> Dict[FieldHeadNames, Tensor]: outputs = {} for field_head in self.field_heads: encoded_dir = self.direction_encoding(ray_samples.frustums.directions) mlp_out = self.mlp_head([encoded_dir, density_embedding], dim=-1)) # type: ignore outputs[field_head.field_head_name] = field_head(mlp_out) return outputs