ПРОМТЫ | ИИ генерации | ПРОДВИЖЕНИЕ - Светлана Наседкина
IT и технологии · 7 августа 2026 г.
Как сделать детектор объекта и перенести туда ваш товар? 💚 Вы можете поменять надписи ... — 7 августа 2026 г. в 15:51:37.154
Как сделать детектор объекта и перенести туда ваш товар? 💚 Вы можете поменять надписи на интерфейсе в промте на более подходящие вам. Если объект получается не того размера: прикрепите к запросу в ИИ третьим фото пример объекта относительно рук и пропишите в промте использовать размер объекта как на 3 фото. 1️⃣ Как всегда, использовала для генерации изображений Nano Banana в SyntX ‼️Я в последнее время работаю через ВЕБ-ВЕРСИЮ ➡️Ссылка на SYNTX.AI в Телеграм https://t.me/syntxaibot?start=aff_1257194318 ➡️Ссылка на SYNTX.AI в браузере https://syntx.ai/welcome/RiJ4FRtC вы можете использовать любой агрегатор или любую нейросеть😉 Промт: Overhead urban street photograph of a young woman walking alone across a zebra crossing on dark asphalt, shot from a high angle as if captured by a surveillance camera, diagonal white crosswalk lines forming strong graphic patterns across the frame. Subject wearing a fitted beige tank top, loose beige wide leg-style jeans, black belt, black flip-flops, black rimless oval sunglasses, holding [REFERENCE OBJECT] in one hand, natural relaxed grip, confident forward stride. CRITICAL RULE: Preserve the subject's facial features exactly 1:1 with the reference photo — face shape, eyes, nose, lips, jawline, skin texture, hair color and style must remain completely unchanged and photorealistic. The face must be instantly recognizable as the same person from the reference image. CRITICAL RULE: If the [REFERENCE OBJECT] contains any text, logos, labels, branding, or inscriptions on its packaging or surface, they must be preserved exactly 1:1 with the reference photo — no distortion, no alteration, no misspelling, no artistic interpretation. Every letter, symbol, color, and placement of text must remain pixel-perfect and legible, identical to the original packaging. Preserve the original photo completely: real face, skin texture, hair, clothing, posture, lighting, street texture, and perspective must remain photorealistic and unchanged. Overlay a futuristic AI surveillance interface (HUD) in fluorescent neon green. Replace the face-detection frame with a rectangular object-detection frame locked onto the [REFERENCE OBJECT] in the subject's hand, with corner brackets and tracking markers targeting the object precisely. Include a profile ID panel on the left side of the frame showing a monochrome image of the [REFERENCE OBJECT] inside a green UI box. Add synthetic system text and labels such as: - "ОБНАРУЖЕНА ОФФЛАЙН-АКТИВНОСТЬ" - "ID: P-2026-05" - "КОЭФФИЦИЕНТ СТИЛЯ: 99%" - "ДОБАВОЧНЫЙ ЭФФЕКТ УВЕРЕННОСТИ В СЕБЕ: ВЫСОКИЙ". Add subtle technical overlays: - thin grid lines - signal strength indicators - tracking dots on the [REFERENCE OBJECT] and the subject's hand holding it - timestamp "May 2026" and system status text - frame borders suggesting a camera feed. All UI elements must be flat, 2D sharp, digital, with clean vector lines, no depth, no glow bloom, no blur. The interface should feel like a government or AI monitoring system, minimalistic, cold, and analytical, with the object as the primary target of surveillance rather than the person. Color palette limited to neon green + white UI lines over a muted, desaturated street photo. Mood: dystopian, controlled, observed, cyber-surveillance aesthetic. Sharp focus, no motion blur, no grain, no painterly effects, no 3D UI, no watermark. Vertical composition, cinematic framing.