Workload-Aware Evaluation of Structured Pruning and Token-Resolution Restructuring for SAM 3 Concept Detection: Segmentation Trade-Offs and Prompt-Count-Aware Accounting

Resource-constrained vision sensor systems must balance promptable concept detection, segmentation quality, and inference cost. We evaluate structured pruning and grid-preserving token-resolution restructuring of pretrained SAM 3 under prompt-count-aware timing. Historical 500-image detection results describe recovered operating points with different training histories. A separate three-arm, three-seed study controls a short 1000-step recovery budget. On all 5000 COCO validation images, native-mask AP is 0.4714 for the baseline, 0.2168 for Grid36-B6 and 0.0831 for Grid24-B4, so these points do not preserve segmentation quality. On this mask metric, the historical pruned checkpoints rank higher (0.3630 and 0.3109), whereas the short matched-recovery control orders the arms the other way. Five independent paired timing sessions distinguish cached-text box calls from uncached-text mask calls. Historical prompt-path optimizations show why trunk acceleration alone does not predict high-prompt-count latency. Label-recovered pruning is sensitive to prompt phrasing. Further studies compare one INT-FUP pooling adaptation, simulated multi-turn image refinement, and native dynamic FP8 in vision MLPs; each retains explicit quality and runtime limits. Five compiled edge artifacts were measured repeatedly on device, but upstream build effects were not isolated; one artifact received unnormalized input, and incomplete matching-artifact accuracy, including a baseline INT8 collapse, prevents accuracy-preserving edge claims. The evidence supports workload-specific detection/runtime trade-offs rather than general compression superiority. Real video and concepts unseen throughout model development remain unvalidated.

Authors

Institutions

Publication Details

Journal
Sensors
Published
2026-11-13
DOI
https://doi.org/10.3390/s26227175
Primary Topic
Advanced Neural Network Applications
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Workload-Aware Evaluation of Structured Pruning and Token-Resolution Restructuring for SAM 3 Concept Detection: Segmentation Trade-Offs and Prompt-Count-Aware Accounting

Kyoungseok Noh, Sung‐Jo Yun, Young‐Ho Choi, Jihyun Park et al.
Sensors
Advanced Neural Network Applications
article

Workload-Aware Evaluation of Structured Pruning and Token-Resolution Restructuring for SAM 3 Concept Detection: Segmentation Trade-Offs and Prompt-Count-Aware Accounting

Kyoungseok Noh, Sung‐Jo Yun, Young‐Ho Choi, Jihyun Park, Jungwoo Lee, Hyogon Kim
article en

Abstract

Resource-constrained vision sensor systems must balance promptable concept detection, segmentation quality, and inference cost. We evaluate structured pruning and grid-preserving token-resolution restructuring of pretrained SAM 3 under prompt-count-aware timing. Historical 500-image detection results describe recovered operating points with different training histories. A separate three-arm, three-seed study controls a short 1000-step recovery budget. On all 5000 COCO validation images, native-mask AP is 0.4714 for the baseline, 0.2168 for Grid36-B6 and 0.0831 for Grid24-B4, so these points do not preserve segmentation quality. On this mask metric, the historical pruned checkpoints rank higher (0.3630 and 0.3109), whereas the short matched-recovery control orders the arms the other way. Five independent paired timing sessions distinguish cached-text box calls from uncached-text mask calls. Historical prompt-path optimizations show why trunk acceleration alone does not predict high-prompt-count latency. Label-recovered pruning is sensitive to prompt phrasing. Further studies compare one INT-FUP pooling adaptation, simulated multi-turn image refinement, and native dynamic FP8 in vision MLPs; each retains explicit quality and runtime limits. Five compiled edge artifacts were measured repeatedly on device, but upstream build effects were not isolated; one artifact received unnormalized input, and incomplete matching-artifact accuracy, including a baseline INT8 collapse, prevents accuracy-preserving edge claims. The evidence supports workload-specific detection/runtime trade-offs rather than general compression superiority. Real video and concepts unseen throughout model development remain unvalidated.

SensorsVol. 26(22)
Korea Institute of Robot and Convergence (KR)
Ministry of Trade, Industry and Energy, Korea Evaluation Institute of Industrial Technology
Advanced Neural Network Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon