Add SDPose pose detection tutorial docs

[lin-bot23]

Summary

Add SDPose + RT-DETRv4 pose detection tutorial documentation (PR #12748).

New files

• tutorials/utility/pose-detection-sdpose.mdx — EN
• zh/tutorials/utility/pose-detection-sdpose.mdx — ZH
• ja/tutorials/utility/pose-detection-sdpose.mdx — JA
• docs.json — registered in all three language sidebars under Utility

Details

SDPose is a whole-body pose detection model natively supported in ComfyUI. The documentation covers:

• Model overview and key strengths (whole-body keypoints, multi-person, configurable outputs)
• Four workflows: Multi-Person (Image), Multi-Person (Video), OOD Image to Pose, OOD Video to Pose Map
• Each workflow with Cloud & Download CardGroup
• Model download: sdpose_wholebody_fp16.safetensors in models/checkpoints/ + rt_detr_v4-x-hgnet in models/diffusion_models/
• Drawing options (body/hands/face/feet toggle) and detection options (threshold, class, max detections)
• Subgraph docs link

Workflow templates:

• utility_sdpose_multi_person.json
• utility_sdpose_multi_person_video.json
• utility_sdpose_ood_image_to_pose.json
• utility_sdpose_ood_video_to_pose_map.json

Model repository: https://huggingface.co/Comfy-Org/SDPose
PR: Comfy-Org/ComfyUI#12748