Q3Computer Graphics and Multimedia

In high-end Computer Aided Design (CAD) and 3D geometric modeling, drawing complex, physically smooth curves cannot be achieved using primitive lines or circles. Engineers demand a highly flexible mathematical architecture where the physical shape of a curve is dictated by a set of interactive "Control Points." The curve is aggressively mathematically drawn toward these control points using complex parametric blending functions. The two absolute foundational architectures in the industry are the Bezier Curve and the B-Spline Curve.

1. The Bezier Curve Architecture

Developed by Pierre Bezier at Renault, the Bezier curve is mathematically defined by a specific set of $N+1$ physical control points ($P_0,P_1,...,P_n$). The curve aggressively begins exactly on the absolute first point $P_0$ and terminates exactly on the absolute last point $P_n$. The intermediate points do not touch the curve; they act as massive magnetic forces, violently pulling the mathematical curve toward them.

• The Mathematical Engine: The exact $(x,y)$ coordinate of the curve at any parametric time $t$ (where $0\le t\le 1$) is calculated using the rigorous Bernstein Polynomials acting as blending functions. $$P(t)=\sum _{i=0}^n{P}_i\cdot B_{i,n}(t)$$
• Convex Hull Property: The entire mathematical curve is strictly guaranteed to be physically contained within the "Convex Hull" (the bounding polygon) formed by connecting the control points. It can never mathematically escape this boundary.
• The Fatal Demerit (Global Control): The Bezier curve suffers from a catastrophic architectural flaw: it only possesses Global Control. If an engineer mathematically moves a single intermediate control point, the entire continuous length of the curve violently alters its shape. This makes fine-tuning highly complex physical models nearly impossible. Additionally, the polynomial degree is strictly bound to the number of control points ($Degree=N$). If you have 100 control points, you must calculate a massive 99th-degree polynomial, destroying CPU efficiency.

2. The B-Spline (Basis Spline) Architecture

To violently solve the catastrophic flaws of the Bezier curve, mathematicians engineered the B-Spline. It completely decouples the mathematical degree of the polynomial from the physical number of control points.

• The Knot Vector: Instead of a single massive polynomial, a B-Spline is mathematically constructed by stitching together multiple low-degree curve segments (like 3rd-degree cubics) at specific junction points called "Knots".
• The Ultimate Merit (Local Control): This segmented architecture grants the B-Spline the holy grail of modeling: Local Mathematical Control. If an engineer moves a single control point in a B-Spline network, only the specific local curve segments immediately adjacent to that point are physically altered. The rest of the massive curve remains absolutely, mathematically frozen and untouched.
• Polynomial Independence: An engineer can aggressively place 500 control points but strictly force the mathematics to remain a highly efficient 3rd-degree (cubic) polynomial, saving massive computational overhead.

Architectural Conclusion

While Bezier curves are mathematically elegant and widely used in 2D vector graphics (like Adobe Illustrator fonts), their catastrophic lack of local control makes them useless for advanced engineering. B-Splines (and their successor, NURBS) dominate all high-end 3D graphics hardware (Maya, AutoCAD) because of their aggressive local control and mathematical stability.