Clothing Image Attribute Point Acquisition via Coordinate Segmentation

The patent segments the attribute points into two categories: five key points that require manual calibration (shoulder angle points, midpoint of lower hem, midpoints of sleeve edges) and eleven points that can be automatically determined through coordinate system calculations. This segmentation allows manual precision for critical points while achieving automation for less critical points, thereby resolving the contradiction between measurement precision and productivity.

The patent applies preliminary action by first manually calibrating the five key attribute points to establish an accurate coordinate system, then using this pre-established system to automatically determine the remaining eleven points. This preliminary manual calibration ensures reliability for subsequent automatic calculations, while significantly reducing the total time required compared to calibrating all points manually.

The patent introduces a coordinate system as an intermediary between manual calibration and automatic point determination. The five manually calibrated points serve as mediators that define the coordinate system, which then mediates the automatic calculation of the remaining eleven points. This intermediary approach enables automation and improves productivity while keeping the system relatively simple, as it only requires establishing a coordinate system rather than implementing complex image processing algorithms.