PV-WAVE Foundation provides a general purpose set of image display and image processing operations. When a more robust set of operations is required the PV-WAVE: Image Processing Toolkit is the right tool for the job. Included in the PV-WAVE: Image Processing Toolkit is an extensive set of filters, transforms and image processing operators designed to meet the needs of even the most demanding image processing application.

Highlights

• Robust graphical user interface makes the PV-WAVE: Image Processing functionality easy to use.
  • Image file import and export
  • Image processing
  • Histograms
  • Profiles
  • Contour Plots
  • Surface Plots
• Import and export of the most common image file formats (BMP, GIF, JPEG, MIFF, PCD, PCX, PNG, SUN, TGA, TIFF, XBM, XPM, XWD). Not all formats are supported on all operating systems.
• Extensive set of point operations and image filters
• Robust set of image processing functions including morphological operations and wavelets
• State of the art classification and segmentation functions
• Full support for Region of Interest (ROI) processing (regular and irregular ROIs)
• Full support for all pixel data types (byte, integer, long, floating point, double precision floating point)
• Support for multi-layered images (apply operations to each plane or a single plane)
• Support for volume processing (3D arrays) and signal processing (1D arrays)
• Batch processing for multiple images (video sequences) or multiple signals (time sequences)

Point operations are image processing operations where each pixel in the output image is dependent only upon the corresponding pixel in the input image. In general, point operations are mathematical and/or logical operations performed on a single image, between a single image and a constant value, or between two images of equal size.

Mathematical Operations

• Addition, Subtraction, Multiplication, Division Blending
• Absolute Value
• Modulo

Logical Operations

• Boolean AND, OR, XOR, NOT

Other Algebraic Operations

• Exponentiation
• Logarithm (natural and Base 10)
• Trigonometric operations

Thresholding

• General thresholding
• Density slicing

Filtering is an image processing operation that is normally used to remove unwanted information in an image or to enhance the information already present.

Image of an MRI slice of a knee (looking from the side, called a saggital slice

Shows a Sobel edge enhancement of the knee image after the noise was removed using our nonlinear geometric mean filter. Now the edges of the bone, etc. are clearly visible.

Shows a Sobel edge enhancement of the knee image. Because there is noise present in the image, the edge image amplifies that noise. The amplified noise drowns out the desired edges.

Edge Detection/Enhancement

• Kernels
  • Sobel, Roberts
  • Frei-Chen, Kirsch, Prewitt
• Operations
  • Canny
  • Shift and subtract
  • Shift and XOR
    

Noise Removal

• Smoothing (Boxcar Average Filtering)
• Median Filtering
• Lee Filtering

Noise Generation

• Impulse, periodic, uniform, Gaussian, Rayleigh, Poisson, exponential

Linear Filters

• Convolution
• Custom spatial filter design
• Save/Restore custom filters

Other Spatial Domain Filters

• Nonlinear Filters
  • alpha-trimmed mean
  • contra-harmonic mean
  • geometric mean
  • maximum, minimum, mode, range
  • rank, Y_p mean
• Adaptive Filters
  • Adaptive double-window-modified trimmed mean (DWMTM)
  • Adaptive minimum mean-squared error filter (MMSE)

Frequency Domain Filters

• Butterworth Filter Design
  • High/Low/Band Pass
• Ideal Filter Design
  • High/Low/Band Pass
  • Notch
• Image Restoration
  • Wiener Filter Design
• Windowing Functions
  • Blackman, Chebyshev, Hamming
  • Hanning, Kaiser, Rectangular
  • Triangular

Morphological Image Processing

Image preprocessing for pattern recognition and image analysis applications often involves morphological operations. Morphology means shape, therefore, morphological image processing routines alter or utilize the shapes within an image.

• Erosion
• Dilation
• Morphologic opening, closing, outlining
• Skeletonization
• Top-Hat transform
• Hit-or-Miss transform

Mensuration

Mensuration refers to the quantification of object features within the image. Mensuration operations are useful in classification and object recognition.

• Shape Description
  • Centroid, 2D Moments
  • Major Axis, Perimeter
• Statistical Measures
  • Maximum, Minimum, Mode, Range
  • Entropy, Kurtosis, Skewness
  • Uniformity
• Other
  • Euclidean distance mapping

Representation and Description

Image representation and description are operations which are sometimes used to preprocess images for pattern recognition and classification. Histogram statistics can provide global image information. Texture analysis can be important in describing regions in an image. Image correlation is often used in template or prototype matching for pattern recognition.

Histogram

• Histogram Representation
  • Density
  • Cumulative density
• Histogram Analysis
  • Histogram statistics

Texture

• Spatial Textural Analysis
  • Gray level co-occurrence matrix (GLCM) and statistics
  • Gray level run length (GLRL) matrix and statistics
• Spectral Textural Analysis
  • Polar FFT




An indexed color image of apples provided by the The University of Chicago Vision and Robotics Group Texture Image Database.	Shows the same image converted to grayscale. PV_WAVE Image Processing can convert between 24 bit color, indexed color, and linear grayscale.
A plot of the summation of the "rho" variable for the polar fft (FFT on a polar coordinate system) of the BW apples image. This variable can give information about the texture of the image which can be used in image classification.	A plot of the summation of the "theta" variable for the polar FFT. Spikes in this signal can give information about the periodicity of the texture.

Correlation

• Direct (spatial domain)
• Indirect (spatial frequency domain)

Image Transforms

There are numerous transforms that can be applied to any image. The two most common transforms are the fast Fourier transform (FFT) and its inverse (IFFT). The FFT converts an image from the spatial domain to the spatial frequency domain. Many other transforms exist and are useful for various applications.

• Power spectrum estimation
• Discrete Cosine (DCT and IDCT)
• Hough (line or circle)
• Principle components (PCT and IPCT)
• Radon
• Slant
• Haar
• Wavelet (including Biorthogonal, Coifman, and Daubechies filter generation)

This is a 24 bit image of a close up of astro turf carpet provided by The University of Chicago Vision and Robotics Group Texture Image Database.	A Hough transform of the image can reveal information about straight lines.

Geometric Transforms

Geometric transforms such as image rotation, scaling, and warping are important in many applications. In particular multi-modal data can be registered to a common coordinate system through the use of geometric transforms. Geometric transforms modify the spatial relationships between image pixels.

• Scaling
• Rotation
• Shifting (Translate)
• Interactive Warping

Color Image Processing

There are two general categories of color image processing: full color and pseudo-color processing. Almost all PV-WAVE: Image Processing Toolkit operations can be applied to binary, grayscale, pseudo-color, and full color images. Exceptions include the morphological processing routines (erosion, dilation, etc. ) and the spatial textural analysis routines (GLRL, GLCM) which are only defined for binary and grayscale images.

Color Models

• 24-bit color to 8-bit color conversion
• 8-bit color to 24-bit color conversion
• Nonlinear grayscale or pseudo-color image to linear grayscale conversion

Image Operations

• Apply image operations to all planes or a single plane of a 24-bit color image

Classification and Segmentation

Segmentation is used to identify regions of common characteristics in an image. Classification is a step beyond segmentation in which particular substances or objects are identified within an image and segmented. Classification usually involves the determination of the number of separate classes contained within the image.

• Supervised classification using the maximum likelihood classifier
• K-Means clustering,
• Region counting, finding, growing, merging, splitting, statistical analysis
• Adaptive thresholding
• General thresholding