AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Perception Characteristics Distance: Measuring Stability and Robustness of Perception System in Dynamic Conditions under a Certain Decision Rule

Boyu Jiang, Liang Shi, Zhengzhi Lin, Lanxin Xiang, Loren Stowe, Feng Guo

CVPR 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Multimodal trajectory prediction · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
rank models by aPCD instead of AP/AR/F1perception benchmark on SensorRainFall · Object detection, vehicle, rainy nightmodel ranking: GLIP aPCD 37.299 vs DyHead 21.546; DyHead AP50:95 0.144 vs GLIP 0.133Mixed
rank models by aPCD instead of AP/AR/F1perception benchmark on SensorRainFall · Instance segmentation, pedestrian, clear daylightmodel ranking: Mask2Former aPCD 23.547 vs ConvNeXt-V2 23.448; ConvNeXt-V2 AP50:95 0.068 vs 0.055Mixed
compare aPCD with traditional metricsperception benchmark on SensorRainFall · 16 experiments (conditions, objects, tasks)model ranking agreement: agreement in 11 out of 16 experimentsMixed
rainy daylight instead of clear daylightGrounding DINO vehicle detection · SensorRainFall, pthres=0.5, ythres=0.5PCDHurt
switch from clear/rainy daylight to rainy night conditionsGrounding DINO vehicle detection · SensorRainFallPCD vs thresholds smoothnessHurt
lower pthres and ythres (0.1-0.3) versus higher (0.6-0.9) (ablation)PCD · SensorRainFallPCDImproved
increase sample size for variance change point detection (ablation)variance change point test · Synthetic simulation, 1,000 replicationsdetected change pointsImproved
increase variance ratio between segments (ablation)variance change point test · Synthetic, 50-50 samplesdetection significance: approximately 3x variance increase sufficientImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy