Presentation on theme: "Forensic Applications of Computer Vision"— Presentation transcript:
1Forensic Applications of Computer Vision Peter KovesiDepartment of Computer Science& Software EngineeringThe University of Western Australia
2Application areas Image enhancement Image metrology Biometric identification
3Video surveillance images are typically of poor quality Image EnhancementVideo surveillance images are typically of poor qualitycameravideo recordermonitorThe quality of the display on the monitor can be misleading.The video recorder runs 24 hours/day. The tape gets worn and the recording heads get dirty. The images that are recorded can be very poor.
4Image Enhancement Common problems in images : Poor resolution, especially in video images.Poor contrast due to under or over-exposure.Corruption with noise.Motion blur or poor focus.Misalignment of rows from line jitter and interlacing of motion in video images.
5Zooming in does not help! Poor resolution:Zooming in does not help!
6Create a `higher resolution’ image by mathematically interpolating pixel values
7Super-resolution from multiple low resolution images Images must bealigned accurately.Object should be planar.Irani and Peleg 1991Capel and Zisserman 1998
8Super-resolution from multiple low resolution images:Results from synthetic data can be impressive
9Real data: Car number plate - low resolution + blur and noise.
10Super-resolution from multiple low resolution images:Number plate reconstruction
11Current method of choice is wavelet shrinkage denoising. Donoho and Johnson 1995original image