Analyzing Statements on 3D Computer Graphics and Perspective Projection

This question asks us to evaluate three statements related to perspective projection and transformation in the context of 3D computer graphics. Let's analyze each statement carefully.

Evaluating Statement (A): Parallel Lines under Perspective Projection

Statement (A) says: Under perspective projection, each set of parallel lines in the object do not stay parallel in the image (except those that are parallel to the viewplane to start with).

Perspective projection is designed to mimic how human vision works, where objects appear smaller as they move further away. This effect is known as foreshortening. A key characteristic of perspective projection is that parallel lines in 3D space, unless they are parallel to the projection plane (or viewplane), will appear to converge at a single point in the 2D image. This point is called a vanishing point.

• If a set of parallel lines is *not* parallel to the viewplane, they will converge to a vanishing point in the projected image, thus not remaining parallel.
• If a set of parallel lines *is* parallel to the viewplane, their vanishing point is at infinity, meaning they will remain parallel in the projected image.

Therefore, statement (A) accurately describes the behavior of parallel lines under perspective projection. It states that parallel lines generally do not stay parallel *except* for the specific case where they are parallel to the viewplane. This statement is correct.

Evaluating Statement (B): Perspective Transformation and the 'z' Coordinate

Statement (B) says: Applying a perspective transformation in the graphics pipeline to a vertex involves dividing by its 'z' coordinate.

In the typical graphics pipeline, a 3D vertex $(x, y, z)$ is transformed through various stages. After being multiplied by the perspective projection matrix, a vertex in homogeneous coordinates $(x', y', z', w')$ is obtained. For perspective projection, the $w'$ component is often related to the original $z$ coordinate. The next step is the perspective divide, where the $x'$, $y'$, and $z'$ components are divided by $w'$. This division is what achieves the perspective foreshortening effect, mapping the view frustum into a canonical view volume (like the Normalized Device Coordinates cube). The depth information, related to the original 'z', is preserved after this division in the new 'z' coordinate, and the original 'z' value influences the 'w' component used for division.

While the division is technically by the $w'$ component after matrix multiplication, this $w'$ is directly derived from the original $z$ and other potential transformations. The fundamental operation that creates the perspective effect and scales coordinates based on depth is this division, often conceptually linked to the original 'z' depth. In many formulations or simplified explanations, the core idea involves scaling $x$ and $y$ inversely with $z$ (or a transformed version of $z$).

Given the context of graphics pipelines, statement (B) captures the essential mechanism: a division step, related to the depth ('z') coordinate, is crucial for perspective transformation. This statement is correct in principle regarding the core operation.

Evaluating Statement (C): Is Perspective Transformation a Linear Transformation?

Statement (C) says: Perspective transformation is a linear transformation.

A linear transformation $L$ satisfies two properties:

• Additivity: $L(v_1 + v_2) = L(v_1) + L(v_2)$
• Homogeneity: $L(cv) = cL(v)$ for any scalar $c$.

Linear transformations can be represented by matrix multiplication. However, as discussed when evaluating statement (B), the perspective transformation process involves a division step (the perspective divide). This division by a coordinate value (which varies per point) is not a linear operation. For example, transforming $(x, y, z)$ to $(\frac{x}{z}, \frac{y}{z}, \dots)$ does not satisfy the properties of a linear transformation.

While the initial multiplication by the perspective projection matrix is a linear operation in homogeneous coordinates, the subsequent perspective divide makes the overall perspective transformation from 3D world coordinates to 2D screen coordinates (after division) non-linear in Cartesian coordinates.

Therefore, statement (C) is incorrect.

Conclusion

Based on the analysis:

• Statement (A) is correct.
• Statement (B) is correct.
• Statement (C) is incorrect.

The statements that are correct are (A) and (B).

Revision Table: Key Concepts in 3D Graphics

Additional Information: Projections and Transformations

In 3D computer graphics, projections are used to map 3D points onto a 2D surface (the screen). There are two main types:

• Orthographic Projection: Parallel lines remain parallel, and objects do not appear smaller with distance. Used in technical drawings and some games. It is a type of affine transformation.
• Perspective Projection: Creates a sense of depth and realism. Parallel lines converge, and object size is inversely proportional to distance.

Transformations in graphics pipeline include model, view, projection, and viewport transformations. The perspective transformation is a crucial step that converts view-space coordinates into clip-space (using matrix multiplication) and then Normalized Device Coordinates (using the perspective divide). This process ensures that objects within the viewing frustum are correctly scaled and positioned for rendering.

Linear transformations can be represented by a $4 \times 4$ matrix multiplication in homogeneous coordinates. Examples include translation, rotation, and scaling. Affine transformations are linear transformations followed by a translation; they preserve parallelism but not necessarily angles or lengths. Perspective transformations involve division and are therefore not affine or linear in Cartesian coordinates.