Structured lighting to create 3D models

Apr 20, 2022by, Suhail Aliyar

Uncategorized

3d Modelling

In computer vision, the process of developing a mathematical representation of any surface of an object in three dimensions via specialized software is called 3D modelling and the product is called a 3D model. A 3D artist is someone who works with 3D models. 3D models can be displayed as a 2D image through 3D rendering or used in a computer simulation of physical phenomena. Most of us know well about 3D Printers, which are the devices that help to simulate these models. At present, animations and 3D modelling are being used in various areas, like the military, automobile engineering, medical science, movies, computer gaming, geology, astrophysics and others.

Common techniques used in 3D modelling

We will discuss the various techniques used for 3-dimensional modelling, focusing on Structured light techniques and application in the field of Orthodontics(Dental Science). The various modelling techniques are introduced below:

Subdivision/Box modelling

Box modelling is a type of polygonal modelling, where one uses a geometric shape, like a cube, cuboid, sphere or cylinder and shapes it until the intended appearance is achieved. Box modellers/Subdivision modellers mainly carry out the process at different levels. They start rolling with a low-resolution grid and then the shape is refined. Then the grid is further subdivided and ensures the hard edges are smoothened, which produces a less pixelated image to reduce noises and include necessary information. The repeated technique of refining and subdividing is processed until enough polygonal details are available into the grid, which conveys the desired concept. Box modelling is combined with edge modelling technique and is one of the most common polygonal modelling techniques used by experts in this area.

Edge/Contour modelling

Contour modelling is one another type of polygonal modelling technique, which is very much different than box modelling. We are familiar with contours as it is the term used to denote object boundary. In the process of edge modelling, the architect develops the model piece by piece, instead of primitive shape refining. This is achieved by placing the loops of many polygons along the outlines or boundaries and filling the areas that lie between them. It is difficult to complete certain meshes through box modelling. For example, the human face cannot be completed only through box modelling.

Spline/NURBS modelling

Spline modelling is extensively used in the automotive and industrial modelling processes. A NURBS is a kind of a mesh which does not have any edge, vertex or face. These models contain smooth surfaces. The concepts are developed by architects by lofting a mesh between splines. Curves developed using a tool, similar to the pen tool used in Adobe Photoshop or MS Paint, are called NURBS curves, which are the primary element in this type of modelling. The architects draw the NURBS curve in the 3D space and modify them by moving the control vertices, which is a series of handles. The curves are placed along prominent contours. The space between the contours is interpolated automatically by the software. Also, a curve can also be created using a profile curve, which revolves around the same in a central axis. This is one of the most used 3D model design techniques used to design objects like glasses, apparels, vases, wine glasses and plates.

Sub-division modelling

Mixing polygonal modelling and NURBS modelling techniques creates sub-division modelling. The 3D models are created through a polygon model as in polygonal modelling and then it is converted into a sub-division model in this process. The architect enjoys control in certain areas over the refinement of the 3D model. These models can be easily transferred by various software available in the market. Until the details are clear enough to display the desired model, the polygon is subdivided many times and refined.If subdivision is greater , then the surfaces will be smoother.

Digital sculpting

The technical industries integrated these various 3D modeling processes and are called disruptive technologies. These models are developed with advanced 3D modeling software. The automobile industry also uses these technologies for making their way into the marketing strategies and  product development. Digital sculpting, a kind of disruptive technology, has been leveraged to a great extent in the 3D modeling process. The architects need not carry out the constraints of edge flow and topography. This enables to design 3D models, in a way similar to the process of digital clay sculpting. Here, meshes are created organically. A tiny device is used for shaping the model, just like  using a brush on chunks of clay by sculptors. Creature and character sculpting has reached new levels with digital sculpting. The architects can carry out the process much faster and with greater efficiency. The architects can work with meshes containing high resolution, over millions of polygons.

Procedural modeling

Procedural modeling refers to designs that are generated by algorithms, and are not created manually by the architects. Here, the objects and scenes are developed by the architects on the basis of user-defined rules. In various other environment modeling packages, the architects create entire landscapes by updating the rules like foliage density and elevation range. They choose from landscapes like deserts, coastal areas or alpines. This modeling is extensively used in organic objects like trees and foliage, where the variations and  complexities are more infinite. These models are not drawn by hand as this is an extremely difficult process to be done by hand. These objects are further tweaked through various custom settings. The density of branches, height of tree trunks, curls and angles, can be editable as per the need.

Image-based modeling

In image-based modeling, from a set of 2D models, 3D objects are derived algorithmically and that are static in nature. This type of modeling is used in cases, when the architects face budgetary or time restrictions. These techniques are not capable of developing fully realized 3D images. Image based modelling is the most common type of 3D modeling in medical science. This is very much familiar across architects or modelers in the entertainment sector as well.

Surface modeling

Surface modeling helps in creating 3D splines. The process incorporates utilizing 2D splines, and is different from NURBS modelling. This technique is used to generate organic 3D models in films and offers a good amount of flexibility to the architects. Various requirements can be easily created by 3D representation, using geometric entities like curves, lines or surfaces.

3D scanning

The 3D scanning technology in real world objects can be digitized, and the process involves high photo-realism levels. The objects and even the actors can be analyzed, scanned and architects can use the raw data to generate NURBS or polygonal mesh using 3D scanning. These contours carry high accuracy and resemble the original figure to a great extent. This process is used when the architects require a digital representation of the Object. Digital representations are extensively used to represent objects, without having any real-world equivalent. The use of 3D modeling in medical science has increased manifold and presently, 3D scanning is one of the most extensively used processes in creating digitized images in the industry of medical science.

 

Use Case : 3D Imaging using Structured Lighting in Orthodontics

Over the past few years, orthodontic and dentofacial orthopedic diagnosis and treatment planning have relied majorly on computer vision and mechanical supports such as imaging, functional analyses and jaw monitoring. The major goal of these techniques are to replicate the physiological and anatomical facts exactly and to visualize the 3D anatomy precisely.

Imaging is considered one of the most important tools for orthodontist treatment and diagnosis in evaluating and recording the size and form of craniofacial structures. Orthodontist experts use 2D static imaging techniques for recording craniofacial anatomy. However, they face difficulty with recording the deepness of structures and localization. 3D imaging has been developed in the earlier days of the 1990’s and has gained a vital place in dentistry, especially in orthodontics, and in orofacial surgical applications. A series of anatomical data is gathered in 3D diagnosis imaging using certain equipment, processed by a computer machine and later visualized using a 2D monitor to present the illusion of depth.

There are a large number of diagnostic methods that have been developed to display facial structures, most of which were abandoned later due to their many cons. 3D imaging techniques is the most popular method of current medicine which gives out  detailed and problem-oriented information about hard and soft tissues, such as Cone Beam Computerized Tomography (CBCT), Computerized Tomography (CT), 3D laser scanning, structured light technique, Micro Computerized Tomography (MCT), 3D surface imaging systems (stereophotogrammetry), 3D facial morphometry (3DFM), Tuned-Aperture Computed Tomography (TACT), and Magnetic Resonance Imaging (MRI) and a lot more.

3D scanning in Medical Imaging

In dentistry, after capturing the face, laser scanners supply 3D images for treatment planning, diagnosis or evaluating effects of orthodontic and orthognathic treatment. The 3D laser scanners produce digital models. Even though it has some cons as with any techniques like, the procedure is so slow , as sometimes distortion can occur during scanning which may affect the scanned image. Also while the scanner revolves around the point of interest(patients’ head), patients should stay motionless for the duration of the scan. Due to the patient’s mobility and security issues related to laser, intraoral laser scanning is considered very difficult to obtain digital models. The safety issues are vital, like exposing eyes to the laser beam, particularly in small children. There are also some inabilities to get soft tissue surface texture, which may result in difficulties in identifying landmarks due to surface pigmentation.

Structured light techniques

The facial aesthetics diagnosis needs to be related to  the deeper structures of muscles and bones, it can be feasible only at the surface level to investigate the face. The structured light scanning technique enables the 3D shape of the face in a simple way without ionizing radiation. The result of this technique is a 3D shape of the patient’s face, visualized on a computer monitor. The 3D facial analysis is accessible and 3D superimposition treatment effects would come into use. The image is illuminated by the light mostly infrared light and taking one single image is very much sufficient in structured light technique. The position of illuminated points is the vital point here,  and the image obtained is necessary for 3D reconstruction of the object. The main aim of this technique is to use the facial shape and underlying radiographic data to construct 3D structures, which assist in diagnosing and evaluating the treatment goals and results.  Also, using the structured light technique in the mouth 3D images of the teeth can be obtained. However, the face needs to be illuminated with random patterns of  high-concentration samples a few times. This increases the increased probability of head action by capture time. In addition, an ear to ear facial model can not be ensured with a single imager. The structured light technique determines the position of the brackets used in dentistry accurately. Ora-Scanner is the first 3D hand-held intraoral scanner and is based on structured light techniques. This system uses white light for illuminating. Kuroda and Techalertpaisan uses 2 LCD projectors, a digital camera and a computer machine, to capture the 3D image of a face that may be moved, altered or revolved lightly in all directions. This system takes at least 2 seconds to capture an image, which may be considered too long in avoiding the movements of babies and children. Another kind of structured light technique is the technique that contains a system with dual cameras and a projector. Before obtaining every image hue coded light is projected. The displacement of the pattern enables the software to evaluate an accurate 3D model by enabling displacement of the patterns. Structured light figure and stereophotogrammetry is united to measure the light figure precisely, which generates accurate 3D plans that help thousands of patients.