SEM I 2025-26, EEE F435 Digital Image Processing
Potential Course Project Topics
Date: 09/19/2025
1. Automated Old Photograph Restoration
This project would focus on taking scanned old, faded, or noisy photographs and automatically improving
their quality.
•
Concepts Used: Image Enhancement (histogram processing for contrast), Image Restoration (noise
models, spatial filtering to remove scratches or grain), and Gray Level Transformations.
2. Simple Optical Character Recognition (OCR)
Develop a program to recognize printed characters from a clean, scanned document. The system would isolate
individual characters and identify them.
•
Concepts Used: Image Segmentation (thresholding to create a binary image), Morphological
Processing (to clean up character shapes), Representation & Description (using boundary descriptors
or PCA to create features for each character), and Object Recognition (using decision-theoretic
methods to classify characters).
3. Traffic Sign Detection and Recognition
Create an application that can detect and identify specific traffic signs (e.g., a stop sign) from an image of a
road.
•
Concepts Used: Image Segmentation (color-based segmentation to find red octagons), Morphological
Image Processing (to refine the shape of the detected sign), and Object Recognition (using pattern
matching to confirm the object is a stop sign).
4. Digital Watermarking in the Frequency Domain
This project involves hiding a small, invisible image or piece of text (the "watermark") within a larger image.
The watermark should be recoverable even if the image is slightly altered.
•
Concepts Used: Fourier Transforms and Filtering in the Frequency Domain (modifying DFT
coefficients to embed the watermark).
5. Medical Image Enhancement for Diagnosis
Focus on enhancing medical images like X-rays or MRIs to make features such as tumors or fractures more
visible to the human eye.
•
Concepts Used: Image Enhancement (using spatial filtering like sharpening), Histogram Processing
(to improve contrast), and Image Segmentation (using edge detection to outline regions of interest).
6. Implementation and Comparison of Image Compression Algorithms
Implement a lossless compression technique like Huffman coding and a basic lossy technique using the
Discrete Fourier Transform (DFT). The project would compare their performance based on compression ratio
and image quality.
•
Concepts Used: Image Compression (Huffman coding, LZW, Run length coding, etc.) and Fourier
Transforms.
7. Automated Defect Detection for Manufacturing
Simulate an industrial inspection system that finds defects (like cracks, holes, or blemishes) on products from
a conveyor belt.
1
•
Concepts Used: Morphological Image Processing (using erosion, dilation, or hit-or-miss
transformation to isolate defects) and Image Segmentation (using thresholding to separate defects
from the background).
8. Document Skew Correction and Cleanup
Create a tool similar to a mobile scanning app that takes a photo of a document taken at an angle, corrects the
perspective, and cleans up the background.
•
Concepts Used: Image Registration (to detect corners and apply a perspective transform), Image
Segmentation (adaptive thresholding to convert the text to binary), and Image Restoration (spatial
filtering to remove noise).
9. Content-Based Image Retrieval (CBIR)
Build a basic system where a user provides an image, and the program retrieves all other "visually similar"
images from a database. Similarity could be based on color, texture, or shape.
•
Concepts Used: Histogram Processing (for color similarity), Representation & Description (using
regional descriptors for texture/shape), and Object Recognition (pattern matching to find the closest
feature sets).
10. Wavelet-Based Image Analysis
This project would explore the use of wavelets for multi-resolution processing. A practical application could
be image denoising or feature detection at different scales.
•
Concepts Used: Wavelets and Multi-resolution Processing (using the wavelet transform to represent
the image at different levels of detail).
2