Image Processing Projects

Collection of image processing projects developed for an Image Processing course that I attended during my undergraduate days at the Instituto Federal Fluminense. Covers classical image analysis techniques implemented from scratch in Java as ImageJ plugins, and a deep learning pipeline built in Python on Google Colab.

Repository Structure

image-processing-projects/
├── README.md
├── java/
│   ├── 01_BloodCellAnalysis/
│   │   ├── BloodAnalysis.java
│   │   └── README.md
│   ├── 02_ColorChannelManipulator/
│   │   ├── SplitChannels.java
│   │   ├── MergeChannels.java
│   │   └── README.md
│   ├── 03_RGB2Grayscale/
│   │   ├── ColorBarGenerator.java
│   │   ├── RGBToGrayscale.java
│   │   └── README.md
│   ├── 04_Point2PointOperations/
│   │   ├── ImageAdjustments.java
│   │   └── README.md
│   ├── 05_HistogramOperations/
│   │   ├── HistogramOperations.java
│   │   └── README.md
│   ├── 06_SpatialFilters/
│   │   ├── SpatialFilters.java
│   │   ├── NonLinearFilters.java
│   │   └── README.md
│   ├── 07_MorphologicalOperations/
│   │   ├── MorphologicalOps.java
│   │   └── README.md
│   ├── 08_RegionOfInterest/
│   │   ├── ROIExtractor.java
│   │   └── README.md
│   └── 09_ConnectedComponents/
│       ├── ConnectedComponents.java
│       └── README.md
└── colab/
    ├── 10_YOLOTrainingPipeline/
    ├── roiExtraction.ipynb
    ├── yoloDatasetGen.ipynb
    ├── yoloTraining.ipynb
    └── README.md

Java Projects (01–09)

All Java projects are ImageJ plugins written and compiled in Eclipse IDE. Each implements its processing pipeline from scratch without relying on ImageJ's built-in filters, except where noted.

#	Project	Topics
01	Blood Cell Analysis	Finding white blood cells in a microscope picture
02	Channel Split and Merge	RGB bit manipulation, packed integer encoding
03a	Color Bar Generator	Synthetic image generation, RGB encoding
03b	RGB to Grayscale	Luminance weighting (BT.601, BT.709), method comparison
04	Image Adjustments	Brightness, contrast, solarization, desaturation, live preview via DialogListener
05	Histogram Operations	Histogram expansion, histogram equalization via CDF
06	Spatial Filters	Linear convolution (mean, sharpen, edge); Sobel operator; median filter
07	Morphological Operations	Erosion, dilation, outline extraction, morphological skeleton
08	Batch ROI Extractor	Batch image processing, Otsu segmentation, particle analysis, file I/O
09	Connected Components	BFS flood-fill labeling, 8-connectivity

General Installation

1. Copy the .java file(s) into the plugins/ folder of your ImageJ installation.
2. Compile via Plugins > Compile and Run, selecting the .java file. Alternatively, compile with javac against ij.jar and drop the .class file into plugins/. The plugin will appear in the Plugins menu after restarting ImageJ.
3. Open an appropriate input image and run the plugin from the Plugins menu.

Input type requirements vary per plugin — check the individual README for each project.

Colab Projects Overview

Project 10 consists of three notebooks designed to run in sequence on Google Colab with a GPU runtime:

1. roiExtraction.ipynb — segments objects from a source image (e.g. a sprite sheet) using Gaussian blur, Otsu thresholding, and contour detection, and exports the cropped ROIs as a ZIP.
2. yoloDatasetGen.ipynb — composites the extracted ROIs onto downloaded background images to generate a synthetic labeled dataset in YOLO format.
3. yoloTraining.ipynb — splits the dataset, applies augmentation, and trains a YOLOv5m detector for 100 epochs.