サーマルビジョンエフェクト

Transform the uploaded image into an AI-enhanced thermal infrared surveillance scan.

Preserve the original composition, camera angle, framing, number of people, body shapes, poses, clothing shapes, hairstyles, facial structure, and recognizable identities. The people must still clearly resemble the original subjects.

Create a stylized but believable thermal vision effect using false-color infrared mapping. Use the full thermal spectrum naturally: deep navy and dark blue for very cold areas, cobalt blue and cyan for cool surfaces, green for mild warmth, yellow and orange for warm body heat, red and small white-hot highlights only for the hottest areas.

Important: preserve facial identity through the thermal effect. Faces should not become featureless heat blobs. Keep recognizable eyes, eyebrows, nose shape, mouth shape, jawline, cheek structure, hairline, and expressions visible as thermal contours. The face should look like a thermal scan enhanced by AI facial recognition.

Thermal heat logic: exposed skin such as faces, hands, necks, ears, and ankles should be the hottest areas, glowing yellow, orange, red, with subtle white-hot highlights. Hair should be cooler than skin, mostly dark blue, purple, or deep red with warm edges near the scalp. Clothing should show body heat underneath but not be solid orange — thin shirts can glow yellow-orange near the torso, knees, elbows and folds; thick hoodies, jackets and shoes should appear cooler with blue, cyan, green and orange heat leakage. Buildings, pavement, sky, glass, trees and distant background stay mostly cold: deep navy, cobalt blue, cyan, dark purple. Windows, lights, engines and electronics may show small realistic yellow-orange heat signatures.

Make the image feel like a high-tech thermal surveillance camera with AI vision enhancement. Add soft thermal bloom around warm bodies, subtle heat bleed, infrared glow, sensor noise, scanlines, digital grain, mild compression artifacts, and slight machine-vision sharpening.

Add clean tactical HUD elements: thin corner frame lines, small targeting reticle, subtle detection boxes around each person, tiny temperature labels, compass markings, thermal scale bar, timestamp-style telemetry, and small scanner readouts. Keep HUD elements secondary and do not cover the faces.

Negative prompt: normal photographic color, true-color photo, featureless thermal faces, melted faces, identity loss, all-orange people, random rainbow filter, inaccurate heat mapping, cyberpunk neon, fantasy glow, anime, cartoon, CGI, plastic skin, changed identity, changed pose, changed clothing, changed background, extra people, missing people, distorted faces, unreadable faces, excessive HUD, text covering faces.