Forensic Image Analysis

The Vision Research Group in the School of Computer Science & Software Engineering has been researching a number of image enhancement techniques suitable for forensic image analysis. Some of these include:

• Image denoising using wavelet techniques.
• Image deblurring using the Richardson-Lucy algorithm.
• Contrast enhancement through greyscale remapping or homomorphic filtering.
• Super resolution from multiple low-resolution images.
• Image rectification

For some details see:

• An overview of forensic image processing techniques - in the form of a seminar presenation.
• Some more detailed information - mainly concentrating on the denoising of images.
• A short PowerPoint presentation given at ANZFSS in Canberra 2002
• Why most surveillance video is useless. Presentation given at ANZFSS 2006
• Why I use MATLAB for forensic image processing

Some interesting links

• House of Lords Science and Technology - Fifth Report Digital Images as Evidence
• FBI Guidelines for digital imaging
• Esther Neate's finger print web site

Return to:

• Peter Kovesi's home page
• The Vision Research Group home page
• The UWA School of Computer Science & Software Engineering home page