Genur.art

"summary": "meta": [ "1024x1536 vertical cover composition", "dynamic high-angle group portrait", "12-degree clockwise camera roll", "all three characters looking directly at the viewer", "main title: MOSS image", "subtitle: 2.5D anime", "typography reference only from the newly attached image" 画面阵列与构图布局 整体阵列：画面采用了非常经典的对角线纵深阵列（阶梯式斜线构图）。以金属栏杆的钢丝为视觉引导线，从左后方向右前方延伸，形成强烈的透视感（近大远小）。 ], "canvas": "single", "caption": "Create a more dynamic version of the three-girl street cover. Preserve the purple-haired central girl, green-haired rear-left girl, blonde right-side girl, their costumes and the original street setting. Move the camera above the group and look downward, with a pronounced but controlled clockwise camera tilt. All three girls lift their faces and look directly into the lens. Preserve their recognizable gestures while giving the hair, ribbons and skirt hems a coordinated sense of movement. Add the main title \"MOSS image\" and subtitle \"2.5D anime\" using the white expressive brush lettering seen in the newly supplied typography reference. Keep \"MOSS image\" together on one line. The newly attached image supplies lettering characteristics only; do not import its five characters, staircase, clothing or wording. No other text, logos, signatures or watermarks." , "items": [ "label": "object", "caption": "Camera and perspective: position the camera above and in front of the group, looking downward approximately 35 degrees. Roll the camera approximately 12 degrees clockwise so the building supports, sidewalk seams and curb share a consistent diagonal orientation. The faces tilt upward toward the lens. Reveal the top surfaces of the hats, shoulders and pavement. Apply moderate wide-angle perspective, making the central foreground girl slightly larger than the two companions without enlarging any individual limb disproportionately. Keep all three faces and all visible shoes inside the frame.", "bbox": [-768, -512, 768, 512] , "label": "character", "caption": "Central purple-haired girl: preserve her long purple hair, red eyes, ornate dark hat with gold bands, black-and-white costume with gold trim and white heeled boots. She raises her chin and looks directly at the camera with an open cheerful smile. One arm bends upward so its hand grips her hat brim. The opposite arm extends toward the blonde girl, with the elbow bent and the hand hanging loosely beside her companion's upper torso. One knee lifts toward the viewer's left beneath the white skirt, while the other leg supports her on the sidewalk. Maintain clear hip, knee and ankle connections. Her long hair and skirt panels swing slightly outward, following one shared breeze.", "bbox": [-355, -345, 650, 250] , "label": "character", "caption": "Rear-left green-haired girl: retain her high green hair tuft, green-and-yellow clothing and green shoes. She leans inward toward the group, lifts her face and looks directly up into the camera with a small lively smile. Shift the hanging black hair ornaments only enough to leave both eyes readable. Her arms stay close to her body and remain partly concealed behind the central figure. Her two legs and shoes remain separate from the central girl's raised knee and white boots.", "bbox": [-440, -440, 515, -70] , "label": "character", "caption": "Right-side blonde girl: preserve her high twin ponytails, white fluffy collar, large red chest bow, brown bodice, rounded pink-and-orange skirt sections, brown leg coverings and yellow boots. Open her eyes and direct both pupils toward the elevated camera while retaining a broad happy smile. Both arms remain raised in peace signs, one above and behind the central hat and one beside her own head. Each hand has two extended fingers and three naturally folded digits. Lean her shoulders slightly inward to tighten the group composition. Keep both boots in contact with the sidewalk and let the ponytails and hanging ribbons lift slightly with the shared breeze.", "bbox": [-405, 85, 645, 465] , "label": "object", "caption": "Anatomy and motion: depict exactly three people, each with one head, two arms and two legs. Preserve the ownership of all gestures: the hat-gripping hand and relaxed hanging hand belong to the purple-haired girl; the two peace-sign hands belong to the blonde girl. The green-haired girl's obscured hands remain concealed. Keep fingers, sleeves and hair clearly separated. Convey movement through body lean, lifted fabric edges and streaming strands while keeping the faces sharp and the feet grounded. Do not use motion blur over faces, hands or lettering.", "bbox": [-440, -440, 650, 465] , "label": "object", "caption": "Main title: write exactly \"MOSS image\" on one continuous line in the open upper wall area, above the characters and beneath the structural beam. Match the supplied lettering reference with bold white brush-drawn forms, a forward slant, energetic sweeping strokes, irregular tapered terminals and restrained broken edges within the strokes. MOSS is uppercase; image is lowercase. Keep the two words together with a clear normal space. Make the main title substantially larger than the subtitle. Add a narrow deep-plum offset shadow for readability against the pale wall, without a rectangular banner. Keep the title inside safe margins and clear of the road sign, raised hands and hair.", "bbox": [-645, -350, -465, 405] , "label": "object", "caption": "Subtitle: write exactly \"2.5D anime\" immediately beneath the main title. Use the same white slanted brush lettering, with smaller strokes and approximately 40 percent of the main title's letter height. Preserve the decimal point and uppercase D. Keep the subtitle on one line, with generous separation from the main title and a matching subtle deep-plum shadow. Give the two-line title block a gentle upward baseline angle similar to the lettering reference, while maintaining easy readability. Do not copy the reference's words \"Style\" or its separate MOSS and Image line breaks.", "bbox": [-450, -230, -365, 340] , "label": "object", "caption": "Environment and layout: retain the pale building facade, dark horizontal and vertical supports, upper windows, white grid-patterned lower wall, partly cropped red-rimmed circular sign at the upper-left, green shrub at the lower-left and pale banded utility pole along the right boundary. Preserve the yellow-tan sidewalk, diagonal curb and gray roadway. Reproject these existing elements consistently for the higher camera and camera roll. Keep the three-character arrangement unchanged: green-haired behind-left, purple-haired in the center foreground and blonde on the right. Reserve the upper wall for the title without adding architecture or relocating the group.", "bbox": [-768, -512, 768, 512] , "label": "object", "caption": "Lighting: maintain warm directional sunlight from the upper-right. Bright highlights fall on the upturned faces, hair crowns, white costume panels and yellow boots. The hat brim creates a controlled shadow above the eyes without hiding their gaze. Soft reflected light from the pavement illuminates the shaded cheeks and lower costume surfaces. Architectural shadows remain coherent across the tilted wall and sidewalk. Keep the background slightly less prominent than the faces and white title, with distinct contact shadows beneath the shoes. Preserve readable eyes and natural highlights without introducing new spotlights.", "bbox": [-768, -512, 768, 512] ], "bbox_coord_space": "abs_centralized", "bbox_format": "yxyx"