Align bbox coordinate handling across primary and fallback paths, normalize Both-mode raw mask rendering, and tighten demo result typing to reduce runtime/display inconsistencies.
Apply Oracle-guided cleanup to make the demo pipeline contract explicit and remove defensive runtime indirection while preserving existing visualization behavior.
Integrate an opt-in OpenCV visualizer into the demo runtime so operators can monitor tracking, segmentation, and inference confidence in real time without changing the default non-visual execution path.
Add preprocess-only silhouette export and configurable result exporters so demo runs can be persisted for offline analysis and reproducible evaluation. Include optional parquet support and CLI visualization dumps while updating tests and tracking notes for the verified pipeline/debug workflow.
Persist orchestration artifacts, including plan definition, progress state, decisions, issues, and learnings gathered during delegated execution and QA gates. This preserves implementation rationale and auditability without coupling documentation snapshots to runtime logic commits.
Introduce focused unit, integration, and NATS-path tests for demo modules, and align assertions with final schema and temporal contracts (window int, seq=30, fill-level ratio). This commit isolates validation logic from runtime changes and provides reproducible QA for pipeline behavior and failure modes.
Add the full demo runtime stack for single-person scoliosis inference, including input adapters, silhouette preprocessing, temporal windowing, ScoNet wrapper, result publishing, and click-based CLI orchestration. This commit captures the executable pipeline behavior independently from tests and planning artifacts for clearer review and rollback.
Add runtime and test dependencies needed by the new demo pipeline, including CLI, typing, model/runtime integration, and messaging libraries. This commit isolates dependency graph updates so environment bootstrap and lockfile drift are tracked separately from feature code.
- ScoNet:
- Previously, `label_ids` remained a NumPy array, which could cause dtype/device mismatches when used with PyTorch tensors on GPU.
- Convert `label_ids` to `torch.from_numpy(...).cuda().long()` to ensure correct tensor type (Long) and device (CUDA), aligning with loss functions that expect class indices on the same device.
- Add new script (`pose_index_extractor.py`) to generate pose selection index for OUMVLP dataset
- Update `pretreatment.py` to load the generated index for pose extraction
crosstyan
Dongyang Jin
zhouzi180
Chao Fan
bugjudger
SHEN CHUANFU
Chengwei Ye
Ahzyuan
Chengwei Ye
Noah