Precise Pose Replication AI Art

The person from image one, with identity and style strictly preserved, re-posed to exactly replicate the full-body pose of the person in image two.

Match image two exactly: subject's position in frame, scale, negative-space ratio, and crop (whether feet/top-of-head/hands are cut). Head tilt, chin up/in, and gaze direction also match image two.

Advanced play: change directional light to side-backlight to instantly make the mood more story-like.

No background-style change that would alter composition; do not add or remove any objects.

Preserve image one's rendering/art style; only the pose and camera composition change to match image two.

Core goal and hard constraints: pose precisely replicated as the sole pose reference (head, neck, shoulders, chest, hips, knees, ankles, arms, wrists, fingers matching in orientation and bend); camera position, shooting angle, lens height, pitch, yaw, perspective and depth of field consistent with image two. Method: treat image two's skeleton as the target skeleton and retarget image one's figure onto it; match center of gravity and pelvis orientation first, then torso rotation and shoulder line, then limb ends (hands, feet), finally fine-tune fingers and expression. Error should be as close to zero as possible. Forbidden: changing clothing, style, or transferring image two's appearance features onto image one.