Pôster de Estilo de Livro de Guia de Companheiro Geométrico

Transform any real photo into a "geometric companion manual style" top-bottom comparison redrawn poster. Composition/framing: strictly split into top and bottom sections - the upper ~45% is a high-fidelity real photograph of the input, and the lower ~55% is a geometric manual-style redraw of the same subject; the horizontal boundary is clear but natural, and side-by-side layouts, nine-grids, and multi-image collages are forbidden. In the lower section, change the subject's scale and information density - do not mechanically trace the original photo composition. Style/medium: paper specification card, field observation page, restrained geometric collage, and hand-printed diagram. Break the subject in the photo into a few flat geometric planes, black outlines of varying thickness, and a limited palette of 2-4 color blocks; preserve the subject's identity, quantity, direction of action, and most critical structure. Use arrows, dashed lines, numbered dots, zoom-in callout boxes, short labels, and clean structural lines to explain visible features, creating an explanatory relationship like "how to observe / how to identify." Annotations must point to real visible parts or spatial relationships of the subject - do not add objects that do not exist in the photo. Scene/backdrop: the lower section uses a warm off-white paper background, with slight paper fiber texture, uneven printing, and low-contrast grain; colors are extracted and compressed from the input photo, not fixed to a blue-sea or dog template. Leave breathing room between the subject and the annotation boxes. Text: include only 1 short English title, 1 material/observation field, and 1 very brief sentence; text should look like labels printed on a paper card - avoid long paragraphs, fake Chinese, URLs, QR codes, brand names, and watermarks. If the photo subject differs, the title and fields must be replaced according to the facts of the photo. Constraints: preserve the subject's identity, quantity, pose, direction of action, key structure, lighting direction, and main spatial relationships from the real photo; the upper section must not be drawn as an illustration, must not replace the subject, and must not blur or reconstruct the photo. The lower section must still clearly correspond to the same subject as the upper section at a glance, and clearly show geometric decomposition, arrow annotations, and the manual-page mechanism. Do not include "original/AI image" corner labels. Avoid: color grading only, grain only, ordinary cartoon filters, realistic photos filling the lower section, meaningless decorative stickers, fixed dog/beach elements, overly complex engineering drawings, dense hard-to-read small text, extra people, and unrelated landmarks.