Dual-Modal Pose Estimation — Live Video + WiFi CSI Fusion
READY
-- FPS ← Dashboard Observatory →
DUAL FUSION
DUAL FUSION

Enable your webcam for live video pose estimation.
Or switch to CSI Only mode for WiFi-based sensing.

◆ Fusion Confidence
Video
0%
CSI
0%
Fused
0%
Cross-modal: 0.000
◆ CSI Amplitude Heatmap
◆ RSSI Signal Strength
-- dBm --
◆ Embedding Space (2D Projection)
◆ RuVector WASM Attention Pipeline
Flash
→
MHA
→
Hyper
→
Linear
→
MoE
→
L+G
Energy: -- Refinement: -- Pose Impact: --
◆ Pipeline Latency
--
Video CNN
--
CSI CNN
--
Fusion
--
Total
◆ Controls
0.30
◆ Live CSI Source
RuView · Dual-Modal Pose Estimation · Architecture: Conv2D → RuVector 6-Stage Attention (Flash+MHA+Hyperbolic+Linear+MoE+L/G) → Fusion → 26-Keypoint Pose
GitHub · CNN: ruvector-cnn (loading…) · Observatory