How Can Polygons Be Clipped? Describe Gouraud Shading Model with Necessary Derivations

---

Part A: Polygon Clipping

Polygon clipping is the process of removing portions of a polygon that lie outside a specified clipping window, while retaining and correctly connecting the portions that lie inside.

Sutherland-Hodgman Algorithm is the most common method for polygon clipping. It clips a polygon against each edge of the clipping window one at a time.

Working Principle:

The polygon (as a list of vertices) is clipped against each of the four boundaries (left, right, top, bottom) in sequence. For each boundary, the algorithm processes every edge of the polygon and produces a new list of vertices.

Four Cases for Each Edge (vertex $S$ to vertex $P$):

• Case 1: Both $S$ and $P$ are inside → Add $P$ to output list
• Case 2: $S$ is inside, $P$ is outside → Add intersection point $I$ to output list
• Case 3: Both $S$ and $P$ are outside → Add nothing
• Case 4: $S$ is outside, $P$ is inside → Add intersection point $I$ and then $P$ to output list

Diagram Description: Imagine a rectangular clipping window with a polygon partially inside and partially outside. The algorithm clips the polygon edge-by-edge against each window boundary, producing a new clipped polygon that fits entirely within the window.

---

Part B: Gouraud Shading Model

Gouraud Shading (also called intensity interpolation shading) is a polygon shading technique that calculates illumination at each vertex and then linearly interpolates the intensity values across the polygon surface.

Purpose: To eliminate intensity discontinuities at polygon boundaries and produce a smooth appearance for curved surfaces approximated by polygon meshes.

Steps in Gouraud Shading:

• Step 1: Determine the average unit normal at each polygon vertex
• Step 2: Apply an illumination model at each vertex to calculate vertex intensity
• Step 3: Linearly interpolate intensities across the polygon surface

---

Derivation:

Step 1: Computing Vertex Normal

The normal at vertex $V$ is the average of normals of all polygons sharing that vertex:

$$\vec{N}_V = \frac{\sum_{k=1}^{n} \vec{N}_k}{|\sum_{k=1}^{n} \vec{N}_k|}$$

where $\vec{N}_k$ are the normals of the $n$ polygons sharing vertex $V$.

Step 2: Intensity at Each Vertex

Using any illumination model (e.g., Phong illumination):

$$I_V = I_a K_a + I_d K_d (\vec{N}_V \cdot \vec{L}) + I_s K_s (\vec{R} \cdot \vec{V})^n$$

This gives intensity values $I_1$, $I_2$, $I_3$ at vertices of a triangle.

Step 3: Interpolation Along Edges

For a scan line at position $y$, interpolate intensity along left and right edges:

$$I_a = I_1 + (I_2 - I_1) \cdot \frac{y - y_1}{y_2 - y_1}$$

$$I_b = I_1 + (I_3 - I_1) \cdot \frac{y - y_1}{y_3 - y_1}$$

where $I_a$ and $I_b$ are intensities at the left and right intersection points of the scan line with the polygon edges.

Step 4: Interpolation Along Scan Line

For a pixel at position $x$ on the scan line:

$$I_p = I_a + (I_b - I_a) \cdot \frac{x - x_a}{x_b - x_a}$$

where $x_a$ and $x_b$ are the x-coordinates of left and right edge intersections.

---

Advantages:

• Eliminates Mach band effect between adjacent polygons
• Computationally less expensive than Phong shading
• Produces smooth shading for diffuse surfaces

Limitations:

• Specular highlights may be averaged out or distorted
• Linear interpolation can miss intensity peaks within polygons

---

Conclusion: Polygon clipping using Sutherland-Hodgman algorithm systematically clips against each window boundary. Gouraud shading provides smooth intensity variation across polygon meshes by interpolating vertex intensities, making it a widely used real-time rendering technique.