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 color transformation commands operate entirely on pixel values — none of them change the spatial layout or dimensions of your image. They range from simple conversions like grayscale and color inversion to more involved operations like Sobel-based edge detection, percentile contrast enhancement, selective RGB channel isolation, and alpha-channel transparency. Each command follows the standard python main.py <command> <src> <dest> syntax, and some pause mid-execution to collect an additional parameter from you interactively.

​
grayscale

The grayscale command converts a colour image to grayscale by computing a luminosity-weighted average of the three colour channels for every pixel. The weighting — R × 0.299 + G × 0.587 + B × 0.114 — matches the ITU-R BT.601 standard and more closely reflects how the human eye perceives brightness than a simple arithmetic mean would.
  • Interactive prompt: none
  • Output suffix: _grayscale
Example 
python main.py grayscale Input_Image/mountains.jpg Output_Image/
Output file: mountains_grayscale.jpg

​
detect-edges

The detect-edges command highlights structural boundaries in an image by applying vertical and horizontal Sobel filters to a grayscale version of the source. For each pixel a 3 × 3 neighbourhood is convolved with each filter; each raw convolution sum is divided by 4 to produce a vertical score and a horizontal score. The final edge score is then computed as: 
edge_score = sqrt(vertical_score² + horizontal_score²)
           = sqrt(v² + h²) / 4
where v and h are the raw convolution sums for the vertical and horizontal filters respectively. The resulting edge map is then normalised to the range [0, 1] by dividing every value by the maximum score in the image, so the output is always a consistent floating-point array regardless of the input image’s intensity range.
  • Interactive prompt: none
  • Output suffix: _edge
Example 
python main.py detect-edges Input_Image/building.jpg Output_Image/
Output file: building_edge.jpg

​
invert-color

The invert-color command produces a photographic negative of the source image by applying a bitwise NOT (~array) to every element of the pixel array. For an 8-bit image this flips each channel value v to 255 - v, turning light areas dark and vice versa.
  • Interactive prompt: none
  • Output suffix: _negative
Example 
python main.py invert-color Input_Image/Who-are-the-hackers.jpg Output_Image/
Output file: Who-are-the-hackers_negative.jpg

​
contrast

The contrast command enhances the perceived contrast of an image using percentile clipping followed by linear rescaling. The process works in three steps:
  1. The 2nd and 98th percentile pixel values of the image are calculated, clipping extreme outliers.
  2. All pixel values are clamped to that range with np.clip().
  3. The clamped values are linearly rescaled to a new maximum defined by your percentage input:
multiplier = int(percentage / 100 × 255)
So a percentage of 80 maps the clipped range to [0, 204].
  • Interactive prompt: Enter contrast percentage: — enter an integer (e.g., 80)
  • Output suffix: _contrast
Example 
python main.py contrast Input_Image/car.jpeg Output_Image/
When prompted, enter your desired percentage: 
Enter contrast percentage: 80
Output file: car_contrast.jpeg

​
rgb-channels

The rgb-channels command isolates one or more colour channels from the image, zeroing out all channels that are not requested. You can request any single channel (r, g, or b) or any combination (rg, rb, gb, rgb). The output is still a three-channel image — unrequested channels are simply filled with zeros.
The input is validated against the regular expression ^r?g?b?$. Any string that contains characters other than r, g, and b — or that arranges them out of order — will be rejected and you will be re-prompted until a valid composition is entered.
  • Interactive prompt: Enter channel composition: — e.g., r, gb, rgb
  • Output suffix: _channel_<composition>
Example 
python main.py rgb-channels Input_Image/mountains.jpg Output_Image/
When prompted, enter your desired channel combination: 
Enter channel composition: rg
Output file: mountains_channel_rg.jpg

​
transparency

The transparency command converts the image to RGBA format and applies a uniform alpha (opacity) value to every pixel. You supply a transparency percentage — the fraction of opacity that will be removed. Internally, the alpha value is calculated as: 
multiplier = int((100 - percentage) / 100 × 255)
So entering 40 means 40 % transparency, leaving the image at 60 % opacity and producing an alpha value of 153 out of 255.
JPEG does not support an alpha channel. Save your output to a PNG file (use a .png destination path) to preserve transparency. If you save the result as a JPEG the alpha channel will be discarded and the transparent areas will appear as solid colour.
  • Interactive prompt: Enter transparency percentage: — float (e.g., 40)
  • Output suffix: _transparency<percentage>
Example 
python main.py transparency Input_Image/peslogo.jpg Output_Image/
When prompted, enter the transparency level: 
Enter transparency percentage: 40
Output file: peslogo_transparency40.0.jpg

Build docs developers (and LLMs) love

Get started for freeTalk to us