---
title: "Image space versus object space methods"
subject: "Computer Graphics"
chapter: 7
key: "semester_5/computer_graphics:image-space-vs-object-space"
source: https://www.bctengineer.com/q/semester_5/computer_graphics/image-space-vs-object-space
site: BCT Engineering
---

# Image space versus object space methods

Computer Graphics · Chapter 7: Visible Surface Determination

- Asked in: 2074 Asoj, 8 marks; 2073 Shrawan, 2 marks; 2069 Chaitra, 3 marks; 2068 Chaitra, 4 marks; 2068 Bhadra (old course), 4 marks; 2067 Mangsir (old course), 4 marks; 2075 Asoj, 4 marks; 2082 Kartik (new course), 1 marks; 2070 Asar, 2 marks (asked 9 times)
- Video: [What is Visible Surface Detection Methods? | Object Space Method | Image Space Method](https://www.youtube.com/watch?v=9agvDVAEU8E) — PiKE Talks
- Answer checked by Claude Opus 5.5 (high effort) on 2026-10-14
- Page: https://www.bctengineer.com/q/semester_5/computer_graphics/image-space-vs-object-space
- Whole chapter: https://www.bctengineer.com/md/semester_5/computer_graphics/chapter-7

## Question

Differentiate between image-space and object-space methods for visible (hidden) surface detection.

## Short answer

Object-space methods compare objects in 3D world coordinates to determine visible surfaces exactly, independent of resolution (e.g., back-face detection). Image-space methods decide visibility per pixel on the projection plane using depth comparisons, resolution dependent (e.g., Z-buffer).

## Answer

Visible-surface detection algorithms are divided into two classes.

- **Object-space methods** compare objects (or parts of objects) with each other in 3D to decide which surfaces are visible. Example: back-face detection, depth sorting (painter's algorithm), BSP tree.
- **Image-space methods** decide visibility pixel by pixel on the projection plane: for each pixel the surface with the nearest depth is found. Example: z-buffer (depth buffer), A-buffer, scan-line method.

| Point | Object-space method | Image-space method |
|---|---|---|
| Works on | Objects/polygons in 3D world | Pixels of the display |
| Accuracy | Precision of the object description (exact) | Limited by screen resolution |
| Basic operation | Compare each object with all others | For each pixel, compare depths of surfaces |
| Cost | Grows with number of objects (about n²) | Grows with number of pixels (resolution) |
| Output | Visible parts of objects, resolution independent | Pixel values for a given resolution |
| Resolution change | Need not be recomputed | Must be recomputed |
| Implementation | Complex geometry, difficult | Simple, hardware friendly |
| Examples | Back-face, painter's, BSP | Z-buffer, A-buffer, scan-line, ray casting |
| Suits | Simple scenes with few objects | Complex scenes |

## Check yourself

The answer's key points as multiple-choice questions, with the correct option and why.

1. Image-space methods decide visibility by working on what?
   A. Pairs of objects compared in 3D
   B. Only the back faces of objects
   C. Each pixel of the projection plane, finding the nearest surface
   D. The edge table of the whole scene only
   Answer: C. Image-space methods work pixel by pixel; object-space methods compare objects in 3D. (Key point: Basis of decision)

2. Which set contains only object-space methods?
   A. Back-face detection, painter's algorithm, BSP tree
   B. Z-buffer, A-buffer, scan-line
   C. Back-face detection, Z-buffer, BSP tree
   D. Painter's algorithm, A-buffer, ray casting
   Answer: A. Z-buffer, A-buffer, scan-line and ray casting are image-space. (Key point: Classification)

3. What happens when the display resolution is changed?
   A. Object-space methods must be recomputed; image-space need not
   B. Image-space methods must be recomputed; object-space results need not be
   C. Both must always be recomputed
   D. Neither is affected
   Answer: B. Object-space output is resolution independent; image-space output is tied to pixel resolution. (Key point: Resolution dependence)

4. Object-space cost grows roughly with what?
   A. The number of pixels
   B. The number of colours
   C. The square of the number of objects
   D. The number of scan lines only
   Answer: C. Each object is compared with all others (about n^2), so many-object scenes are costly. (Key point: Cost growth)

## Related in Computer Graphics

- [Scan-line method of visible surface](https://www.bctengineer.com/question.md?key=semester_5%2Fcomputer_graphics%3Ascan-line-visible-surface) — Computer Graphics, chapter 7, asked 3×
- [Back-face detection algorithm](https://www.bctengineer.com/question.md?key=semester_5%2Fcomputer_graphics%3Aback-face-detection-method) — Computer Graphics, chapter 7, asked 8×
- [A-buffer method](https://www.bctengineer.com/question.md?key=semester_5%2Fcomputer_graphics%3Aa-buffer-method) — Computer Graphics, chapter 7, asked 2×

## Related in other subjects

- [MRI components, image detection, MRI vs CT](https://www.bctengineer.com/question.md?key=semester_7%2Fbiomedical_instrumentation%3Amri-components-vs-ct) — Biomedical Instrumentation, chapter 5
- [Principal points, object for infinity image](https://www.bctengineer.com/question.md?key=semester_1%2Fengineering_physics%3Atwo-identical-lenses-infinity-object) — Engineering Physics, chapter 5
- [BFSK generation, non-coherent detection and signal space](https://www.bctengineer.com/question.md?key=semester_6%2Fcommunication_systems%3Abfsk-noncoherent-signal-space) — Communication Systems, chapter 7
- [Image versus graphics](https://www.bctengineer.com/question.md?key=semester_8%2Fmultimedia_system%3Aimage-vs-graphics) — Multimedia System, chapter 3, asked 2×