GB2246689A - Creating a 3-D image from a plurality of cross-sections - Google Patents

Abstract

In a computer graphics system a 3-D image of an object is created by assembling, or "stocking", a plurality of cross-sections of the object. The cross-sections may be produced by repeatedly sectioning the object and photographing the exposed surface or by the use of an X-ray device such as a CT scanner. Areas of interest in the image can be defined using different types of contour lines and different colours. The image processing involves converting raster-based data to a vector format. The object may be a biological specimen or a man-made object.

Description

SCAN DATA TRANSFER This invention relates to an original and novel way of acquiring and processing computer data to reconstruct as a three dimensional image when an existing object needs to be spatially recorded.

At present to model computer generated three dimensional images the operator has to enter in sets of measurements and instructions based on complex calculations of the actual object by physical and electronicmeasurement devices, with or without sets of existing drawings.

The process is very time consuming and is not necessarily very accurate especially where complex internal structures within the object are needed to be recorded.

According to this invention serial section images are used of the subject. The hard copy whether in positive or negative form on paper, film or any other dense medium is then converted from raster to vector format using existing conversion technology within a computer.

Grey scale or coloured images are converted initially into contour lines of slightly varying types initially prior to this conversion. Different line types are assigned to identify different organs or densities in the material. Line information of this type is capable of conversion to a vector format in this technology.

The two dimensional line images are then produced in a suitable language that will run on a computer that has a 3D modelling software program. The operator simply isolates each image and stacks them around a common datum point recreating the orignal scale between the slices using a common axis.

The number of original serial sections determines the resolution or amount of information in the reconstructed model. Computer techniques may then be used to join up the contours by generating a surface or plane between points of information stored in the contours according to their type of line. The original subject is then recreated in an intelligent or vector format with which transformatiof may be made; such as adding colour, light or shade, rotations, viewpoints, multiple viewing points or assigning values for stress analysis.

For example, one hundred serial sections of an ant may be obtained by using an existing technique based on freezing tissue or immersion in paraffin wax and obtaining micron slices. The images are arranged evenly on a sheet of paper or film each with a common datum point on each section. The images are composed of a range of half tones or grey tones and the dark outline of the dense tissue or bone like structures.

Contour lines of varying width or types (i.e. dotted or dashed) are assigned to differentiate between the scales of grey and/or colour. This may be done by using existing edge tracing techniques on computer graphic systems or simply by enlarging the images and tracing them off by hand and then reducing them to overlay the original dark lines of the bones or dense material image.

The resulting sheet of images then become a collection of contour lines with a common datum or reference point to each image.

The images are now capable of being known by a computer as a series of vectors (or short lines) by using existing conversion technology from raster (the original serial sections or contours) to vector. As a collection of lines of different sorts the image is known in the software as a set of mathematical co-ordinates and thus becomes intelligent. A program may be included to translate dark silhouettes of various width to edge outlines only.

A tape or disc is made in a suitable language (using a suitable translation software) to run on a Computer Aided Design station. The collection of images will then appear on a computer screen. The operator will then isolate each serial section image in the x, y, plane and using a common datum point stack each image in register in the z plane. Surfaces may be defined between the contours of similar types to model volumes.

The ant has been recreated in 3D showing the internal and external proportions. To enhance the image, colour values and/or shades may be assigned to fill those areas bounded by a certain type of width of line. This will differentiate in 3D between the different biological organs within the ant which are indentified by the original grey scale or colour images.

Complex spatial arrangements and structures may then be studied and effectively shown. The resolution of the image dependent upon the number of original slices of the ant, the thickness of the slice and the clarity of the original grey scale or colour image.

Another example would be a complex plastic moulding.

Serial section can be obtained through the use of narrow beam X rays such as are used in medical diagnostics.

These images would be produced as positive images on a sheet of transparent film, in this case there would be no half tones.

The sheets of images would be processed from raster to vector format using existing technology and a translation mode using a suitable software (ie. I.G.E.S.) to recreate the images on a C A D screen. Wide silhouettes would be reduced to line outlines at the conversion stage.

The operator stacks the image on screen reproducing to scale the original intervals around the common axis.

The resulting image is given a surface by joining planes between the contours.

Stress analysis programs may then be run on the accurately defined model in 3D.

Another example would be a human head. Serial sections may be obtained through slicing an existing cadaver at Imm intervals and making single frame photographic images at each point. Alternatively CT scans may be made at regular intervals through a live subject.

Each image would have the organs defined by either colour or grey scale recognition techniques, with or without expert human editing.

The contours are processed from raster to vector format and the slices re-assembled by the C A D operator.

Spatial boundaries and surface planes are generated indentifying different organs or densities in 3 dimensional volumes. Colour values and/or light and shade may be added. Prints, computer plots, videos picture files or films may be made from transformation from the 3D data.

Another example would be a human cell sections which would be obtained through electron microscopy. The images would be processed as previous described.

Images of any subject may be reconstructed depending upon the ability to obtain its serial section.

Claims (8)

1. The exclusive use of serial section images in hard copy form on paper or film taken at regular or varying intervals or intervals that join to each other depending on the width of the serial section; as a source of data for modelling subjects in 3 dimensions on computer systems.

2. The identification of complete structures in 3D within divergent specimens by assigning contour lines of varying types to indentify different types of densities or biological organs or structures.

3. The exclusive use of raster to vector image processing technology with the original source of data as serial section images to precipitate rapid and accurate generation of intelligent computer held 3 dimensional models.

4. Stacking serial section data on computer and generating surface boundaries to define structures or biological organs as articulate spatial volumes.

5. Rapid identification of complex areas as in Claim 2 and Claim 4 by assigning colour and/or shade values to the surfaces joined between the common types of contour line.

6. Rapid editing of the images as in Claim 5 by changing colour and/or shade values relating to these defined areas.

7. Any 3 dimensional computer held model which has been generated by recourse to obtaining serial sections of the existing object and rapid processing by the application of raster to vector technology upon hard copy images either on paper or film or any dense medium or as written software.

8. Indentification of complex areas as in Claim 2 and Claim 4 by labelling with reference to two dimensional and three dimensional images.

Published 1992 at The Patent Office. Concept House. Cardiff Road. Newpon. Owent NP9 IRH. Further copies may be obtained from Sales Branch. Unit 6. Nine Mile Point. Cwmfelinfach. Cross Keys. Newport. NPI 7HZ. Printed by Multiplex techniques ltd. St Man Crap. I