Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Adarsh275/Image-Transformation/llms.txt

Use this file to discover all available pages before exploring further.

The following examples walk through every command in Image Transformation using the sample images bundled in the Input_Image/ directory of the repository. Each example shows the exact command to run, any interactive prompts you will encounter, and the name of the output file that will appear in Output_Image/. Make sure you have activated your virtual environment and that the Output_Image/ directory already exists before running any command.
Commands that require a scaling factor, angle, percentage, or channel composition cannot be driven non-interactively — they always pause and wait for keyboard input. These are: upscale, downscale, rotate, contrast, rgb-channels, and transparency.

​
Color Operations

​
Grayscale Conversion

Convert mountains.jpg to a luminance-weighted grayscale image. The tool applies the ITU-R BT.601 weighting (R × 0.299, G × 0.587, B × 0.114) to each pixel. 
python main.py grayscale Input_Image/mountains.jpg Output_Image/
# Output: Output_Image/mountains_grayscale.jpg

​
Edge Detection

Detect edges in building.jpg using a Sobel filter. The image is first converted to grayscale internally, then both horizontal and vertical gradients are computed and combined. 
python main.py detect-edges Input_Image/building.jpg Output_Image/
# Output: Output_Image/building_edge.jpg

​
Invert Colors (Negative)

Produce a photographic negative of Who-are-the-hackers.jpg by flipping every channel value with a bitwise NOT (~). 
python main.py invert-color Input_Image/Who-are-the-hackers.jpg Output_Image/
# Output: Output_Image/Who-are-the-hackers_negative.jpg

​
Contrast Enhancement

Stretch the pixel value range of car.jpeg by clipping to the 2nd–98th percentile and remapping to a 0–multiplier range. The multiplier is derived from the percentage you enter.
1

Run the command

python main.py contrast Input_Image/car.jpeg Output_Image/
2

Enter the contrast percentage at the prompt

Enter contrast percentage: 80
Entering 80 maps the stretched values onto a 0–204 range (80% of 255).
3

Collect your output

# Output: Output_Image/car_contrast.jpeg

​
RGB Channel Isolation

Zero out one or more colour channels in mountains.jpg, keeping only the channels you specify. For example, entering rg retains the red and green channels while zeroing blue — producing a yellow-toned result.
1

Run the command

python main.py rgb-channels Input_Image/mountains.jpg Output_Image/
2

Enter the channel composition at the prompt

Input should be a combination of r g and b
Enter channel composition: rg
Valid values are any combination of r, g, and b written in that order: r, g, b, rg, rb, gb, or rgb. Uppercase letters and spaces are rejected and will re-prompt.
3

Collect your output

# Output: Output_Image/mountains_channel_rg.jpg

​
Transparency

Add an alpha channel to peslogo.jpg by setting each pixel’s opacity to (100 - percentage) / 100 × 255. Entering 40 means 40% transparent, so 60% opaque.
1

Run the command

python main.py transparency Input_Image/peslogo.jpg Output_Image/
2

Enter the transparency percentage at the prompt

Enter transparency percentage: 40
3

Collect your output

# Output: Output_Image/peslogo_transparency40.0.jpg
Because peslogo.jpg is a JPEG, the alpha channel written by the tool will be discarded when Matplotlib saves the file back as JPEG. To preserve transparency, use a PNG source image and ensure the output path (derived from the source filename) also has a .png extension. See Image Format Considerations for details.

​
Geometric Operations

​
Horizontal Flip

Mirror mountains.jpg left-to-right using NumPy’s np.fliplr(). 
python main.py flip Input_Image/mountains.jpg Output_Image/
# Output: Output_Image/mountains_flipped.jpg

​
Rotation

Rotate building.jpg by an arbitrary angle. The output canvas is automatically expanded to fit the full rotated image, and any unfilled pixels are set to black (zero).
1

Run the command

python main.py rotate Input_Image/building.jpg Output_Image/
2

Enter the rotation angle at the prompt

Enter angle in degrees: 45
The angle must be an integer. The rotation is counter-clockwise.
3

Collect your output

# Output: Output_Image/building_rotated_45.jpg

​
Scaling Operations

​
Upscale

Enlarge peslogo.jpg by repeating each pixel along both axes (np.repeat). Entering 2 doubles the width and height.
1

Run the command

python main.py upscale Input_Image/peslogo.jpg Output_Image/
2

Enter the scaling factor at the prompt

Enter scaling factor: 2
The factor must be a positive integer. A factor of 3 triples each dimension.
3

Collect your output

# Output: Output_Image/peslogo_upscaled.jpg

​
Downscale

Shrink building.jpg by sampling every Nth pixel along both axes. Entering 2 halves the width and height.
1

Run the command

python main.py downscale Input_Image/building.jpg Output_Image/
2

Enter the downscale factor at the prompt

Enter downscale factor: 2
The factor must be a positive integer greater than 1.
3

Collect your output

# Output: Output_Image/building_downscaled.jpg

​
Using the Python API Directly

You can import the imageEdit class directly and chain operations in a script without going through the CLI. This is useful when you want to process many images in a loop or apply several transformations to the same source file.The example below applies grayscale and edge detection to mountains.jpg, saving each result as a separate file:
from src.imageEdit import imageEdit

# Grayscale conversion
gray = imageEdit("Input_Image/mountains.jpg", "Output_Image/")
gray.loadImg()
gray.grayscale()
gray.writeImg("grayscale")
# Saves: Output_Image/mountains_grayscale.jpg

# Edge detection (use a fresh object — writeImg mutates self.dest)
edges = imageEdit("Input_Image/mountains.jpg", "Output_Image/")
edges.loadImg()
edges.edgeDetection()
edges.writeImg("edge")
# Saves: Output_Image/mountains_edge.jpg
Always create a separate imageEdit object for each output file. The writeImg() method mutates self.dest in place — it appends the filename to the destination path, so calling it twice on the same object would produce an incorrect path on the second call.
The same pattern applies to any other method. For example, to upscale programmatically without an interactive prompt, you can set the internal state directly:
from src.imageEdit import imageEdit

obj = imageEdit("Input_Image/peslogo.jpg", "Output_Image/")
obj.loadImg()

# Replicate what upscale() does, but without the input() call
import numpy as np
f = 3  # 3× upscale
obj.outp = obj.inp.repeat(f, axis=0).repeat(f, axis=1)

obj.writeImg("upscaled")
# Saves: Output_Image/peslogo_upscaled.jpg

Build docs developers (and LLMs) love

Get started for freeTalk to us