Human (Input)
hpm_000103.jpg
Garment
top_f_0091.jpg

Thinking Process

PHASE 1: VISUAL DECONSTRUCTION - **Fabric Physics:** The source garment is a structured, houndstooth-patterned vest, likely composed of a wool-blend or tweed material. It possesses high flexural rigidity (stiffness) and will not drape like the user's current silk/rayon shirt. The surface is matte with a textured weave; light reflection should be diffuse with minimal specular highlights, emphasizing the tactile coarseness of the fabric rather than a sheen. - **Key Features:** The notched lapel structure must retain its 3D architectural shape and not flatten against the chest. The distinct black-and-white houndstooth pattern must remain consistent without warping, except where natural folds occur. The welt pocket flaps and the black inner lining visible at the collar are critical anchor points for realism. The sleeveless armholes must be cleanly defined. PHASE 2: FIT & DRAPE SIMULATION - **Body Mapping:** Unlike the user's currently tucked-in shirt, this vest is a longline cut (tunic length). The simulation must untuck the silhouette, allowing the vest to hang straight down past the waistline, covering the top portion of the denim skirt. The garment should rest heavily on the shoulders and hang vertically due to gravity, maintaining a "boxy" fit rather than clinging to the waist curves. - **Occlusion Handling:** The user has long dark hair draping over both shoulders and the upper chest. This is a critical occlusion layer; the hair must be masked and composited *over* the lapels and shoulders of the new vest. The user's hands are positioned low; if the vest length reaches the hands, the fabric must pass behind the thumbs/fingers. PHASE 3: ENVIRONMENT & LIGHTING - **Light Match:** The user environment features soft, diffuse indoor lighting, likely from an overhead source slightly to the front. The vest requires soft ambient occlusion mapping in the lapel folds and under the pocket flaps. Shadows should be soft-edged, projecting slightly downward. - **Color Calibration:** The source image has cooler, studio-like lighting. The white sections of the houndstooth pattern must be warmed slightly (shifted toward a creamy beige) to match the warm/yellow undertones of the user's room lighting and skin tone, ensuring the black-and-white contrast doesn't look artificially stark. PHASE 4: IDENTITY PROTECTION - **Facial Lock:** The entire head region, specifically the eyes, nose, mouth, jawline, and the specific volume and fall of the hair, must be frozen and masked from alteration. - **Body Integrity:** The user's shoulder width and hip alignment must remain constant. Do not slim the torso to fit the vest; instead, warp the vest to accommodate the user's natural measurements. The exposed skin tone on the arms (which will now be visible due to the sleeveless nature of the vest) must perfectly match the user's facial skin tone.
Result
train_0044
Shortcuts
← Prev Image
→ Next Image
Shift+S Search Training ID
Shift+F Search Filename
Shift+H Toggle Help