引導式測驗 · 中等 等級

事件相機與神經形態視覺測驗

測試非同步亮度事件、時間戳記和極性、快速運動優勢、靜態場景限制、閃爍雜訊和感測器融合。

問題 1 的 8

What typically triggers a pixel in an event camera?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. What typically triggers a pixel in an event camera?
    • A sufficient local brightness change
    • Every scheduled frame exposure
    • An object label emitted by a classifier
    • Any object motion regardless of pixel brightness
  2. Which fields usually describe one brightness-change event?
    • A complete RGB frame and exposure setting
    • Pixel position, timestamp and polarity
    • Only a 3D object pose
    • Only global camera luminance without pixel location
  3. Why can event sensing help with very fast motion?
    • It measures every object velocity directly
    • It guarantees zero algorithmic processing delay
    • It prevents all sensor noise during motion
    • It reports local changes without waiting for the next full frame
  4. A stationary object sits in steady light. What limitation can appear?
    • Steady pixels must emit at maximum rate
    • It automatically records a complete color image
    • Few events may describe its static texture or color
    • Polarity identifies the object’s material
  5. Why can flickering lights complicate event-based motion tracking?
    • They synchronize all fused sensors automatically
    • They suppress every noise event
    • They can trigger brightness events without object motion
    • They turn changes into absolute depth measurements
  6. What can conventional frames add when fused with event data?
    • Appearance information missing from a sparse or static event stream
    • Perfect timestamp alignment without calibration
    • A complete 3D map from one exposure
    • Removal of all event-sensor noise
  7. How can an algorithm discard an event camera’s timing advantage?
    • By preserving each individual timestamp
    • By using short windows suited to the task
    • By synchronizing the clocks of fused sensors
    • By pooling events over a very long window before acting
  8. Which evaluation best tests a robot’s event-camera pipeline?
    • Only the advertised microsecond timestamp precision
    • Task error and latency across quiet, flickering and fast-motion scenes
    • Only total event count without task results
    • Only one bright static benchmark scene