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
- Kyoungseok Noh (ORCID: https://orcid.org/0000-0002-0457-2992)
- Sung‐Jo Yun (ORCID: https://orcid.org/0000-0002-4120-9797)
- Young‐Ho Choi (ORCID: https://orcid.org/0000-0003-1576-8277)
- Jihyun Park (ORCID: https://orcid.org/0000-0001-6856-9912)
- Jungwoo Lee (ORCID: https://orcid.org/0000-0001-9091-1729)
- Hyogon Kim (ORCID: https://orcid.org/0000-0002-4765-6117)
Institutions
- Korea Institute of Robot and Convergence (KR)
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
- Ministry of Trade, Industry and Energy
- Korea Evaluation Institute of Industrial Technology