Image Analysis & Processing at Avantier

July 18, 2023

Most images nowadays are recorded in digital form. From a mathematical point of view, they are just 2-dimensional or multidimensional arrays containing values of pixel signals. In simplest binary images each pixel can have only 2 values: zero and one. Gray-scale (black and white, monochrome) or indexed images have one value per pixel, and the value usually is 8-bit or 16-bit, while any other number of bits also may be used. In addition to the image array, color indexed images also have a color table, which assigns at least 3 (Red, Green and Blue) color values to each image array value. More common RGB color images just contain Red, Green and Blue intensity values (typically 8-bit ones) for each pixel. 

Multispectral/Hyperspectral images may contain any number of intensity values per pixel obtained in different spectral ranges. These values often have 16 bits, but may have any number of bits, and may be recorded as float or double values.

Uncompressed image arrays usually are saved in BMP or TIFF or RAW files. Industrial and scientific images normally are not compressed so they preserve all the image details. Consumer quality images commonly are recorded as highly compressed JPEG files, which discard a lot of almost non-noticeable information contained in RAW images.

Medical imaging uses specialized file formats, such as DICOM. Say, MRI or CT scan results are saved in DICOM files as a sequence of many gray-scale 2-dimensional arrays.

Computer generated images, unlike natural images, do not have noise, and so are very suitable for lossless compression in PNG files.

At Avantier Inc., our experienced team of engineers with over 30 years of Image Analysis and Image Processing can assist to image process all files described above and include all types of color image processing for color images, as well as all kinds of image enhancement.  When new Spectroscopic X-ray images, promising to replace MRI and CT, will become available, we will also have the capability for image processing of those as well.

Video files fundamentally also contain just sequences of 2-dimensional pictures. Consumer-grade MPEG files use prediction in time to discard information not noticed by viewers to immensely (100 times or so) decrease movie file sizes. We also have the capability to image extract and image process frames from any kind of video files.

As images are nothing but multidimensional arrays, any information which can be presented as such an array also can be subjected to our processing algorithms. For example, in 3-dimensional cameras, such as Microsoft Kinect, each pixel measures distance to the image point in 3D XYZ space. Kinect saves just a 2-dimensional rectangular array of distance values. Kinect uses additional camera-geometry information to calculate 3-dimensional point cloud XYZ coordinates from the 2-dimensional distance array. We can provide image analysis and image processing for both 2D arrays of distances and 3D point clouds.

Image Analysis

Images are obtained by some image acquisition hardware devices. We take original images, output by such devices, analyze them, and process them according to the analysis conclusions.

Artificial intelligence (AI) approach to building mathematical models is based on using billions of images collected on the Web to train the models to adjust billions of model parameters for image classification. The result is creation of probabilistic models, which sometimes, but not always, provide reasonable conclusions.

Our engineers can analyze big sets of images and find common features of such sets, but to also get more reliable results than pure AI, our in-house experts has the ability to verify claims obtained from the computers.

Our experienced team of engineers uses standard analysis tools such as PCA, histograms, Fourier Transform (FFT), entropy evaluation, etc. To detect hidden images, such as those used in steganography, we analyze all the bit planes of the images. We can also analyze compressed versions of the images, such as Discrete Cosine Transformed (DCT) or wavelet files.

Sponsored Recommendations

Request a free Micro 3D Printed sample part

April 11, 2024
The best way to understand the part quality we can achieve is by seeing it first-hand. Request a free 3D printed high-precision sample part.

How to Tune Servo Systems: The Basics

April 10, 2024
Learn how to tune a servo system using frequency-based tools to meet system specifications by watching our webinar!

Motion Scan and Data Collection Methods for Electro-Optic System Testing

April 10, 2024
Learn how different scanning patterns and approaches can be used in measuring an electro-optic sensor performance, by reading our whitepaper here!

How Precision Motion Systems are Shaping the Future of Semiconductor Manufacturing

March 28, 2024
This article highlights the pivotal role precision motion systems play in supporting the latest semiconductor manufacturing trends.