TotalSegmentator v2 — Worker pipeline v2 (task 297, 118 classes)

All compute (WASM + WebGPU) runs in a dedicated Web Worker — page stays fully responsive during the entire inference. Whole-body CT at 3 mm.  ·  ← v1 (main-thread)
✓ 100.0000% byte-exact canonical-argmax conformance with nnUNetv2_predict CPU (task 297). Uses symmetric pad_nd_image + GPU-side sliding_unpad_accum kernel so postprocess_gpu's per-channel logit resample sees the correct (unpadded) canonical shape.

Log

Axial slice viewer

Canonical input
Segmentation
Overlay