Drawing on images (lines, rectangles, circles, text) in Computer Vision - Deep Dive

Overview - Drawing on images (lines, rectangles, circles, text)

What is it?

Drawing on images means adding shapes like lines, rectangles, circles, or text onto pictures. This helps highlight or mark important parts of an image. It is done by changing the image pixels to show these new elements. Anyone can use this to make images clearer or more informative.

Why it matters

Without the ability to draw on images, it would be hard to explain or point out details in pictures, especially in fields like medicine, security, or self-driving cars. Drawing helps humans and machines understand images better by adding extra information visually. It makes communication with images easier and more effective.

Where it fits

Before learning to draw on images, you should understand basic image formats and how images are represented as arrays of pixels. After this, you can learn about image processing techniques and how drawing helps in tasks like object detection or annotation.

Mental Model

Core Idea

Drawing on images means changing pixel colors to add shapes or text that highlight or explain parts of the picture.

Think of it like...

It's like using a transparent sheet over a photo and drawing with markers to point out or decorate things you want to show.

Image (pixels) ──▶ Modify pixels ──▶ Add shapes/text
  │                     │
  │                     └─ Lines, Rectangles, Circles, Text
  └─ Original picture    └─ Highlighted or annotated image

Build-Up - 6 Steps

1

Foundation

Understanding Image Pixels

Concept: Images are made of tiny dots called pixels, each with color values.

An image is like a grid of pixels. Each pixel has color information, usually in red, green, and blue parts. When you draw, you change some pixels to new colors to create shapes or text.

Result

You see a picture made of many colored dots that can be changed to add drawings.

Understanding pixels is key because drawing means changing pixel colors in specific places.

2

Foundation

Basic Drawing Functions Overview

Concept: Drawing uses simple functions to add lines, rectangles, circles, or text by changing pixels.

Functions like draw_line, draw_rectangle, draw_circle, and draw_text take coordinates and colors to add shapes or words on images. These functions calculate which pixels to change.

Result

You can add simple shapes or words on an image by specifying where and what color.

Knowing these basic functions lets you start marking images to highlight or explain parts.

3

Intermediate

Drawing Lines and Rectangles Precisely

🤔Before reading on: do you think drawing a line changes pixels only on the exact path or also nearby pixels? Commit to your answer.

Concept: Lines and rectangles are drawn by changing pixels along calculated paths or areas.

To draw a line, the function finds all pixels between two points and colors them. Rectangles are drawn by coloring pixels along four edges or filling inside. Thickness controls how many pixels wide the shape is.

Result

Lines and rectangles appear exactly where specified, with chosen thickness and color.

Understanding pixel paths helps control drawing accuracy and appearance.

4

Intermediate

Adding Circles and Text to Images

🤔Before reading on: do you think text drawing changes pixels one by one or uses a font pattern? Commit to your answer.

Concept: Circles are drawn by coloring pixels at a fixed distance from a center; text uses font shapes mapped to pixels.

Circle drawing calculates pixels forming a round shape using radius and center. Text drawing uses font data to know which pixels form each letter, then colors those pixels on the image.

Result

You see smooth circles and readable text added to the image.

Knowing how fonts map to pixels explains why text looks clear or blurry.

5

Advanced

Handling Colors and Transparency

🤔Before reading on: do you think drawing always replaces pixel colors or can blend with existing colors? Commit to your answer.

Concept: Drawing can replace pixel colors or blend them using transparency (alpha) for smooth effects.

Colors have red, green, blue, and sometimes alpha (transparency). When drawing with transparency, new colors mix with old pixels instead of replacing them, creating see-through effects.

Result

Drawings can be solid or semi-transparent, blending naturally with the image.

Understanding color blending allows creating more natural and professional annotations.

6

Expert

Optimizing Drawing for Large Images

🤔Before reading on: do you think drawing on large images is always slow or can be made faster? Commit to your answer.

Concept: Efficient drawing uses algorithms and memory tricks to speed up pixel changes on big images.

Instead of changing pixels one by one, optimized methods update only needed areas, use fast math for shapes, and avoid redrawing unchanged parts. This saves time and memory.

Result

Drawing on large images becomes fast and smooth, suitable for real-time applications.

Knowing optimization techniques is crucial for using drawing in real-world, high-performance systems.

Under the Hood

Drawing changes image pixels by calculating which pixels belong to the shape or text and then updating their color values. For lines, algorithms like Bresenham's find pixel paths. Circles use midpoint or trigonometric calculations. Text uses font bitmaps or vector outlines mapped to pixels. Transparency blends new colors with old ones using alpha compositing formulas.

Why designed this way?

These methods were chosen for efficiency and accuracy. Algorithms like Bresenham's minimize calculations and produce smooth lines. Using pixel-level control allows drawing on any image format. Transparency blending was added to support layered graphics and natural annotations. Alternatives like vector graphics exist but are less direct for pixel images.

┌─────────────┐
│ Input Image │
└─────┬───────┘
      │
      ▼
┌─────────────┐
│ Drawing API │
│ - Line      │
│ - Rectangle │
│ - Circle    │
│ - Text      │
└─────┬───────┘
      │ calculates pixels
      ▼
┌─────────────┐
│ Pixel Buffer│
│ (color data)│
└─────┬───────┘
      │ updates pixels
      ▼
┌─────────────┐
│ Output Image│
└─────────────┘

Myth Busters - 4 Common Misconceptions

Quick: Does drawing text on an image always keep the original image pixels underneath? Commit yes or no.

Common Belief: Drawing text just adds letters without changing the original image pixels beneath.

Reality: Drawing text changes the pixels where letters appear, replacing or blending with original pixels.

Why it matters: Assuming original pixels stay can cause confusion when text overlaps important image details or when trying to remove text later.

Quick: Is drawing a line on an image always a perfect straight line with no gaps? Commit yes or no.

Common Belief: Drawing a line always produces a perfect, gapless straight line on the image.

Reality: Due to pixel grids, lines are approximated and may appear jagged or have gaps depending on thickness and algorithm.

Why it matters: Expecting perfect lines can lead to disappointment and misunderstanding of image resolution limits.

Quick: Does drawing with transparency always look the same on all images? Commit yes or no.

Common Belief: Transparent drawing looks identical regardless of the image underneath.

Reality: Transparency blends with underlying colors, so the final look depends on the original image colors.

Why it matters: Ignoring this can cause unexpected visual results, especially on complex or colorful backgrounds.

Quick: Can drawing on images be done without changing the original image data? Commit yes or no.

Common Belief: You can draw on images without altering the original image pixels.

Reality: Drawing modifies pixel data; to keep the original, you must copy the image first.

Why it matters: Not copying images before drawing can cause loss of original data, making undo or reuse impossible.

Expert Zone

1

Drawing order matters: later drawings can cover earlier ones, affecting visibility and layering.

2

Anti-aliasing smooths edges by blending colors on shape borders, improving visual quality but increasing computation.

3

Font rendering can use vector outlines or bitmap fonts, each with tradeoffs in scalability and clarity.

When NOT to use

Drawing on images is not ideal for scalable graphics or animations where vector graphics or specialized libraries like SVG or OpenGL are better. For complex text layout or multi-language support, dedicated text rendering engines outperform simple drawing functions.

Production Patterns

In real systems, drawing is used for annotating images in medical scans, marking detected objects in self-driving car cameras, or adding captions in social media apps. Efficient layering and caching are used to avoid redrawing entire images repeatedly.

Connections

Image Segmentation

Drawing highlights or marks regions identified by segmentation algorithms.

Knowing how to draw on images helps visualize and verify segmentation results clearly.

Human-Computer Interaction (HCI)

Drawing on images enables interactive tools where users annotate or edit pictures.

Understanding drawing basics supports building intuitive interfaces for image editing or labeling.

Cartography

Both involve overlaying shapes and text on maps or images to convey information.

Recognizing this connection shows how drawing principles apply across visual communication fields.

Common Pitfalls

#1 Drawing directly on the original image without copying it.

Wrong approach:

image = load_image('photo.jpg')
draw_line(image, start, end, color)
# original image pixels changed permanently

Correct approach:

image = load_image('photo.jpg')
copied_image = image.copy()
draw_line(copied_image, start, end, color)
# original image preserved

Root cause:

Not realizing drawing modifies pixels in place, risking loss of original data.

#2 Using integer coordinates without considering image boundaries.

Wrong approach:

draw_rectangle(image, (-10, 20), (100, 200), color)
# negative coordinates cause errors or unexpected results

Correct approach:

x1, y1 = max(0, -10), 20
x2, y2 = min(image.width-1, 100), min(image.height-1, 200)
draw_rectangle(image, (x1, y1), (x2, y2), color)

Root cause:

Ignoring that pixel coordinates must be within image size limits.

#3 Drawing text without specifying font size or color clearly.

Wrong approach:

draw_text(image, (50, 50), 'Hello')
# text may be too small or invisible

Correct approach:

draw_text(image, (50, 50), 'Hello', font_size=24, color=(255,255,255))

Root cause:

Assuming default parameters are suitable for all images and contexts.

Key Takeaways

Drawing on images means changing pixel colors to add shapes or text that explain or highlight parts of the picture.

Basic drawing functions use coordinates and colors to modify pixels for lines, rectangles, circles, and text.

Transparency and color blending allow smooth, natural-looking annotations instead of harsh overlays.

Efficient drawing on large images requires optimized algorithms to avoid slowdowns in real applications.

Understanding pixel-level drawing helps in many fields like image annotation, computer vision, and interactive graphics.

Practice

1. Which OpenCV function is used to draw a rectangle on an image?

easy

A. cv2.line

B. cv2.rectangle

C. cv2.circle

D. cv2.putText

Solution

1. Step 1: Understand drawing functions in OpenCV

   OpenCV provides specific functions for different shapes: cv2.line for lines, cv2.circle for circles, cv2.rectangle for rectangles, and cv2.putText for text.

2. Step 2: Identify the function for rectangles

   The function named cv2.rectangle is designed to draw rectangles on images.

3. Final Answer:

   cv2.rectangle -> Option B

4. Quick Check:

   Rectangle drawing = cv2.rectangle [OK]

Hint: Rectangle drawing uses cv2.rectangle function [OK]

Common Mistakes:

• Confusing cv2.line with rectangle drawing
• Using cv2.circle for rectangles
• Trying to draw text with cv2.rectangle

2. Which parameter in cv2.putText controls the thickness of the text?

easy

A. thickness

B. fontScale

C. fontFace

D. color

Solution

1. Step 1: Review cv2.putText parameters

   The function cv2.putText has parameters: fontFace (font style), fontScale (size), color (text color), and thickness (line thickness of text).

2. Step 2: Identify thickness parameter

   The thickness parameter controls how bold or thick the text lines appear.

3. Final Answer:

   thickness -> Option A

4. Quick Check:

   Text thickness = thickness parameter [OK]

Hint: Thickness of text is set by 'thickness' parameter [OK]

Common Mistakes:

• Confusing fontScale with thickness
• Using color to control thickness
• Mistaking fontFace for thickness

3. What will be the color of the line drawn by this code snippet?

cv2.line(img, (10, 10), (100, 10), (0, 0, 255), 2)

medium

A. Red

B. Green

C. Blue

D. Black

Solution

1. Step 1: Understand BGR color format in OpenCV

   OpenCV uses BGR order for colors, so (0, 0, 255) means Blue=0, Green=0, Red=255.

2. Step 2: Identify the color from the tuple

   Since only the last value (Red) is 255, the line color will be bright red.

3. Final Answer:

   Red -> Option A

4. Quick Check:

   BGR (0,0,255) = Red [OK]

Hint: Remember OpenCV colors are BGR, last value 255 means Red [OK]

Common Mistakes:

• Assuming (0,0,255) is blue (RGB confusion)
• Mixing up color order
• Ignoring OpenCV's BGR format

4. Identify the error in this code that tries to draw a circle:

cv2.circle(img, (50, 50), -10, (255, 0, 0), 3)

medium

A. Center coordinates must be floats

B. Color tuple is wrong format

C. Thickness cannot be 3

D. Negative radius is invalid

Solution

1. Step 1: Check circle parameters

   The radius parameter must be a positive integer representing the circle size.

2. Step 2: Identify invalid radius

   The radius given is -10, which is invalid and will cause an error.

3. Final Answer:

   Negative radius is invalid -> Option D

4. Quick Check:

   Radius must be positive integer [OK]

Hint: Radius must be positive; negative values cause errors [OK]

Common Mistakes:

• Using negative radius values
• Thinking thickness 3 is invalid
• Assuming center coordinates must be floats

5. You want to draw a blue rectangle with thickness 4 and label it "Object" in white text above it. Which code snippet correctly does this?

img = cv2.imread('image.jpg')
start = (30, 30)
end = (150, 150)
# Options below

hard

A. cv2.rectangle(img, start, end, (255, 0, 0), 2)
cv2.putText(img, 'Object', (30, 20), cv2.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 4)

B. cv2.rectangle(img, start, end, (0, 0, 255), 4)
cv2.putText(img, 'Object', (30, 20), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 0, 0), 2)

C. cv2.rectangle(img, start, end, (255, 0, 0), 4)
cv2.putText(img, 'Object', (30, 20), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 255), 2)

D. cv2.rectangle(img, start, end, (0, 255, 0), 4)
cv2.putText(img, 'Object', (30, 20), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 255), 2)

Solution

1. Step 1: Identify blue color in BGR

   Blue in BGR is (255, 0, 0), so rectangle color must be (255, 0, 0) with thickness 4.

2. Step 2: Check text color and position

   Text "Object" should be white (255, 255, 255) and placed above rectangle at (30, 20) with reasonable font scale and thickness.

3. Final Answer:

   Blue rectangle with thickness 4 and white "Object" text above -> Option C

4. Quick Check:

   Blue = (255,0,0), white text, thickness 4 [OK]

Hint: Blue is (255,0,0); white text is (255,255,255) [OK]

Common Mistakes:

• Mixing up BGR color order
• Using wrong thickness values
• Placing text inside rectangle instead of above