Image Processing for Computer Vision

It is a program to perform simple image manipulation with following requirements:

• Load an image by reading it from a file or capturing it directly.
• Image should be read as a 3 channel color image.
• Program should work on any size of image.
• On pressing a support key, apply corresponding function on the original image.
• When implementing convolution, use Cython to speed up execution.

Program was implemented using Python, Cython, and OpenCV. Refer the report for further implementation details and instructions to run the code: View Report

List of Support Keys:

• i - Reload the original image (cancel any previous processing).
• w - Save the current image into the file 'out.jpg'.
• g - Convert the image to grayscale using the openCV conversion function.
• G - Convert the image to grayscale using implemented conversion function.
• c - Cycle through the color channels of the image showing a different channel every time the key is pressed.
• s - Convert the image to grayscale and smooth it using the openCV function. Use track bar to control the amount of smoothing.
• S - Convert the image to grayscale and smooth it using implemented function which should perform convolution with a suitable filter. Use track bar to control the amount of smoothing. Convolution function is implemented from scratch using Cython.
• d - Downsample the image by a factor of 2 without smoothing.
• D - Downsample the image by a factor of 2 with smoothing.
• x - Convert the image to grayscale and perform convolution with an x derivative filter. Normalize the obtained values to the range [0,255].
• y - Convert the image to grayscale and perform convolution with a y derivative filter. Normalize the obtained values to the range [0,255].
• m - Show the magnitude of the gradient normalized to the range [0,255]. The gradient is computed based on the x and y derivative of the image.
• p - Convert the image to grayscale and plot the gradient vectors of the image every N pixels. Length of vector is a fixed value K. Use track bar to control N.
• r - Convert the image to grayscale and rotate it using an angle of theta degrees. Use track bar to control theta.
• h - Display support keys

Results

1. Grayscale image with OpenCV and conversion function:

OpenCV	Conversion Filter

2. Cycle through color channels of the image:

Red Channel	Green Channel	Blue Channel

3. Smoothing image with OpenCV and convolution function:

OpenCV	Cython Function

4. Downsampling with and without smoothing:

Downsampling without smoothing	Downsampling with smoothing

5. X & Y derivative and Magnitude of gradient:

X derivative	Y derivative	Magnitude of gradient

7. Plot gradient vectors of image every N pixels:

Plotted gradient vectors at every 50 pixels

8. Roatation of image:

Rotated image at 45 degrees