Medical-Image-Processing__Morphology-DenoisingMethods-NLM-TotalVariation

This repository contains Python code for various tasks related to medical image analysis and processing. Below are the detailed descriptions and implementations for each section.

Section 1: Morphology

The goal of this section is to create a mask for the man in the picture and process it accordingly.

• Step 1: Determine the range of clothing pixels.
  • Read the image q1.png.
  • Identify clothing pixels based on the green channel.
  • Save the resulting mask as q1res01.jpg.
• Step 2: Cover image holes and create a uniform mask for clothes.
  • Use closing and opening methods to fill holes and smooth the mask.
  • Save the uniform mask as q1res02.jpg.
• Step 3: Redden all areas belonging to the person's clothes.
  • Utilize the obtained mask to identify clothes area.
  • Redden the clothes area in the original image and save as q1res03.jpg.

Section 2: Denoising Methods

This section focuses on denoising techniques applied to medical images.

• Step 1: Gaussian noise addition and grayscale conversion.
  • Add Gaussian noise to the image hand.jpg.
  • Convert the color image to grayscale.
  • Display noisy and clean images side by side.
• Step 2: Classical Regression Filtering.
  • Explain Gaussian kernel formulation and its role in denoising.
  • Remove noise using the Gaussian kernel.
• Step 3: Bilateral Filtering.
  • Explain Bilateral Filtering and its parameters.
  • Remove noise using Bilateral Filtering.
  • Discuss the conceptual meaning of parameters hx and hg.

Section 3: NLM

This section involves noise removal using the Non-Local Means (NLM) algorithm and Gaussian filter.

• Step 1: Apply noise to images and clean using NLM.
  • Add Gaussian and pepper-salt noises to images.
  • Clean the images using the NLM algorithm with varying parameters.
  • Plot PSNR graphs for different filter parameters.
• Step 2: Compare NLM with Gaussian filter.
  • Denoise images using Gaussian filter with different parameters.
  • Calculate PSNR between denoised and original images.
  • Compare results and discuss findings.

Section 4: Total Variation

This section implements the total variation filter for noise reduction.

• Step 1: Create a modified Shepp-Logan phantom with Gaussian noise.
  • Generate clean and noisy images.
• Step 2: Remove noise using total variation filter.
  • Implement the total variation filter with specified parameters.
  • Display clean images, noisy images, and noise removal results.
• Step 3: Calculate SNR criterion for the filter.

Section 5: Total Variation Reproduction

This section reproduces the denoising operation using total variation.

• Step 1: Implement the provided codes for TV denoising.
• Step 2: Apply denoising operation on the noisy image.
• Step 3: Analyze and report results according to the provided description.

For detailed implementation and results, refer to the respective Python files in this repository.