I P P y

Image Processing Pipeline, Desktop Software Application
mac - win - unix - linux

Project Goal & User Story:

Enables the user to perform advanced Digital Signal Processing functions on an image by selecting from a variety of industrial grade default and customizable filters.
In version 1.0 release, that pipeline can be used to single or batch process subsequent images and also carry out Feature Extraction and Pattern Recognition, using Machine Learning algorithms.
Conversion of image to a 3D pixel Elevation Map for alternate view!

Download Beta Version (Free)

Download Link: IPPy_Beta_20170909.jar
Size: 6.2 MB

Written entirely in Java 8, JavaFX 2.2, runnable with a JVM installed on user's OS.


Facial Feature Detection

The Body on Somerton Beach
Filter Pipeline: SRC -> FFD
Mortuary photo of the unknown man found dead on Somerton Beach, south of Adelaide, Australia, in December 1948. Sixty-three years later, the man's identity remains a mystery
Original Image Credit: http://www.smithsonianmag.com/history/the-body-on-somerton-beach-50795611/

Scientific - Astronomy

NGC 7009 (Saturn Nebula) Planetary Nebula; Hubble Space Telescope
Filter Pipeline: SRC -> NEG
Original Image Credit: https://en.wikipedia.org/wiki/Saturn_Nebula#/media/File:NGC_7009_Hubble.jpg
Permission: PD-HUBBLE

Manufacturing - Computer Vision

Simplification of complicated images while retaining important information.
Edge detection filter used for automated analysis of inventory or other tasks.
Machine Learning algorithms can now be applied with higher accuracy.
Filter Pipeline: SRC -> GRY -> CED -> NEG
Canny Edge Detector, sigma=1.4, size=6, lowThreshold=10, highThreshold=20

Geology - Volcanoes

Volcanic Ash Plume arising from Mount Cleveland on May 23, 2006
Seen from the International Space Station.
Filter Pipeline: SRC -> GRY -> SMO-> SMG -> HEM -> NEG
Credit for original Image Used:
Original Image Credit: https://en.wikipedia.org/wiki/Mount_Cleveland_(Alaska)#/media/File:MtCleveland_ISS013-E-24184.jpg
"NASA material is not protected by copyright unless noted".


Contact person for this project can be reached at: email: lance@lancedooley.com