Member Avatar for Samyx

Hello everyone,

I am a MS Computer Science student, 2nd year.

I am looking for ideas of mini image processing projects. I need to come up with a project for my Image Processing class, but the project needs to be done within 4 weeks.

Any suggestion on something that could be interesting, relatively new, and that can be done in a short amount of time?

Thanks

  • 4 Contributors
  • 7 Replies
  • 200 Views
  • 2 Weeks Discussion Span
  • Latest Post Latest Post by sufyan2011

Dani AI

Generated

A short, practical plan that fits a 4‑week window and builds on what you already have (): you mentioned convolution, interpolation, segmentation and histogram work, and rightly suggested starting with a command‑line pipeline then adding a GUI if time allows. Below are three compact, application‑oriented projects that are doable in Java or C++ (OpenCV bindings are a good fit), plus concrete steps and ways to add a small original contribution.

  1. Lightweight face‑attribute classifier (gender / coarse age): detect faces, align/crop, extract simple descriptors (LBP or HOG), train a linear SVM, and evaluate with k‑fold cross‑validation and a confusion matrix. Novelty: produce an error‑analysis view in the GUI that shows misclassified examples and the feature vectors for those cases. Week plan: week 1 dataset + pipeline, week 2 features + training, week 3 tuning + evaluation, week 4 UI + report.

  2. Content‑aware portrait auto‑crop and local exposure fix: build an app feature that scores candidate crops by face inclusion and saliency coverage, then applies local contrast-limited histogram equalization (CLAHE) only to dark regions. Novelty: fast crop scoring and a user slider that trades composition vs. background retention. This is very visual and demo‑friendly.

  3. Small, perceptual compressor for icons/screenshots: implement k‑means color quantization + run‑length + simple entropy coding, compare size/quality to PNG, and add a perceptual quantizer that preserves edges. Novelty: automatic mode that chooses palette size to hit a target file size with minimal visible artifacts.

Scope tips: pick one project, implement a clear baseline and one improvement, keep datasets small (a few hundred to a few thousand images), and start with a command‑line tool before a minimal GUI. Watch for dataset imbalance, lighting variability (preprocess with normalization), and overfitting. Include a short writeup with quantitative metrics (accuracy, confusion matrix, PSNR/visual examples) and a demo—those make the “contribution” clear.

Recommended Answers

All 7 Replies

Member Avatar for Azmah

Do you mean you are making an image manipulation program like Photoshop and/or you want projects to make in Photoshop or something?

Member Avatar for peter_budo

Given you did not specified what "Image Processing", plus you did not mentioned programming language you want/can use it is hard to advise.
However you may consider to write up something like this , (I'm trying to write it at the moment with Java 7 as file metadata is new feature of Java latest release)

Member Avatar for Samyx

The language can be Java or C++, I am just now learning Matlab, so I don't think I will be using that one for this project.
I need a project that will be able to do image manipulation like Photoshop, but it would be something more like on the application side. For example, using what I learnt in my image processing class to do facial detection and differentiate male from female...

The only problem is that the project needs to be done within a 4 weeks period...

Member Avatar for peter_budo

The language can be Java or C++, I am just now learning Matlab, so I don't think I will be using that one for this project.
I need a project that will be able to do image manipulation like Photoshop, but it would be something more like on the application side. For example, using what I learnt in my image processing class to do facial detection and differentiate male from female...

The only problem is that the project needs to be done within a 4 weeks period...

We do not know what you learn so far and given the nature of the project I would say it is very short period to accomplish. Given what you just told us, start with simple console application where you provide image folder as argument at start, do your chosen processing and then provide some form of output/report. Once this accomplished and there is some time remaining try to provide some simple GUI like simple where user can immediately browse for and submit folder location and showing report in some more structured way.

Member Avatar for Samyx

I have already done some application like the one you suggested. I already created an application that will allow the user to select an image and do operations like convolution, interpolation, linear segmentation, histogram and histogram equalization. But it was a homework assignment. I need something a little more advanced. And the project is supposed to provide some kind of contribution, something that is new...

Is there anything new in image compression that I can do? or something simple in face detection?

Member Avatar for peter_budo

Sorry not my area of interest so I wouldn't know...

Member Avatar for sufyan2011

You can do simple face recongnition/detection, the famous image processing project. Though there is a great work in these areas but you can put something new. PCA (principle component analysis) is usually used for face recoginition and skin color tone, edge detection, neure processing etc are used for face detection. It is not big task to develop them in Matlab because toolkits are available in Matlab and you just need to call the methods. You can choose C# for development, this would help you to better understand how to perform image processing in depth like edge detection, guassian blur, mean, median and average etc

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.