ガイド連動クイズ · ハード レベル

ニューラル ラディアンス フィールド クイズ

ポーズをとったトレーニング画像や密度と外観の分離など、オリジナルの NeRF のニューラル シーン表現とレイ レンダリングを確認します。

関連するガイドパス神経放射フィールド
質問 1 の 8

Which task does the original NeRF representation support from its posed scene images?

すべての質問をプレビューする8 質問

クイズに答えると、答えと解説が表示されます。

  1. Which task does the original NeRF representation support from its posed scene images?
    • A guarantee of complete hidden geometry from any two photos
    • Synthesis of views not identical to the training cameras
    • Translation of every scene into a text caption
    • Mandatory face identification for every image
  2. Which quantities are predicted by the original NeRF field for queried scene coordinates?
    • RGB appearance and volume density
    • Only one complete final image returned by every point query
    • The camera’s required GPS address
    • Only the training image filename
  3. What does the conceptual five-dimensional input in the original NeRF paper combine?
    • Five separate input photographs
    • Five independent color channels
    • One date and four document labels
    • A 3D position and a viewing direction with two angular degrees of freedom
  4. How does original NeRF rendering use queried samples along a camera ray?
    • It takes only the brightest sample regardless of every other contribution
    • It ignores sample distances and every density value
    • It composites density and color with sample distances and transmittance
    • It selects an unchanged training image solely by filename
  5. What purpose does positional encoding serve in the original NeRF experiments?
    • Helping the MLP represent high-frequency coordinate-dependent detail
    • Guaranteeing a complete scene from no observations
    • Changing all spatial coordinates into camera GPS readings
    • Eliminating every optimization step
  6. オリジナルの NeRF の実質的な大きな制限は何でしたか?
    • 音声しか出力できなかった
    • 各シーンのトレーニングとレンダリングに時間がかかりました
    • 写真は一切必要ありませんでした
    • 屋内でのみ動作しました
  7. Which camera information is supplied with images in the stated original NeRF training setup?
    • Only a universal claim that all photo collections share one pose
    • Only a text prompt with no image-camera information
    • Known camera poses
    • A required list of real-world product prices
  8. In the original NeRF architecture, how is density’s dependence on input coordinates restricted relative to color?
    • Density must depend only on viewing direction and ignore position
    • Both outputs must be constant across all positions
    • Color must ignore both position and direction
    • Density uses position only; color can also use viewing direction