A fuzzy controller-based framework dynamically selects video model scales using frame correlations and device resources to balance inference accuracy and efficiency.
Experimental results show that the resourc e utilization efficiency index is significantly superior to th at of traditional single-model inference methods
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Breaking the accuracy-resource dilemma: a lightweight adaptive video inference enhancement
A fuzzy controller-based framework dynamically selects video model scales using frame correlations and device resources to balance inference accuracy and efficiency.