Integral imaging is an autostereoscopic or multiscopic 3D display, meaning that it displays a 3D image without the use of special glasses on the part of the viewer. It achieves this by placing an array of microlenses (similar to a lenticular lens) in front of the image, where each lens looks different depending on viewing angle. Thus rather than displaying a 2D image that looks the same from every direction, it reproduces a 4D light field, creating stereo images that exhibit parallax when the viewer moves.
The concept was proposed in 1908 by Gabriel Lippmann, and to date has found use largely in the related concept of lenticular printing of static images.
Contents |
On March 3, 1908, physicist Professor Gabriel Lippmann proposed the use of an array of lenses placed at a picture surface to form true three dimensional pictures. He submitted this proposal to the French Academy of Sciences under the title “La Photographie Integrale”.[1]
The first in-depth study of lithographic printing of integral imagery was developed in 1936 by Carl Percy and Ernest Draper of the Perser Corporation. The first integral animation effect printing was proposed in 1958 by Juan Luis Ossoinak, of Argentina.
A number of researchers continued to advance the process of Integral Photography over the last 40 years, including:
Creating 3-D integral imagery, by digitally interlacing a set of computer generated two-dimensional views, was first demonstrated in 1978 at the Tokyo Institute of Technology in Japan. They and others also developed experimental integral television methods. Digitally interlacing integral imagery for high-resolution color pictures was first proposed in 1989 by Ivars Villums. Many thousands of experimental images have been produced throughout the last century, by a wide variety of methods, exhibiting 3-D, animation and other effects. Research and commercialization of integral methods remains very active today including a wide body of work in integral television and other electronic displays. Integral imaging has not yet achieved significant commercial success, in part because diamond tooling molds for large plastic arrays is difficult, and in most cases prohibitively expensive. Inexpensive lens array sheets have been produced using extrusion embossing (in the same manner as lenticular sheets are made), using anilox patterned rolls, but the resulting lens array pattern, while adequate for moiré effects, is not suitable for integral imagery due to its lack of predictable geometry.
In 2006 an experimental prototype demonstrated the use of a holographic lens array to enhance the viewing angle and DOF of integral imaging systems.[2]
An integral image consists of a large number of closely packed, distinct micro-images, that are viewed by an observer through an array of spherical convex lenses, one lens for every micro-image. The term “Integral” comes from the integration of all the micro images into a complete three dimensional image through the lens array. This special type of lens array is known as a fly's-eye or integral lens array; see Fig. 1.
When properly practiced, the result is credible three-dimensional imagery which coveys a realism similar to holograms. Indeed, it has been demonstrated that an integral image can very accurately reproduce the wavefront that emanated from the original photographed or computer-generated subject, much like a hologram, but without the need for lasers to create the image; see Fig. 2. This allows the eyes to accommodate (focus) on foreground and background elements, something not possible with lenticular or barrier strip methods.
Integral imaging is based on a principle known as the lens “sampling effect”. To achieve this effect, the thickness of the lens array sheet is chosen so that parallel incoming light rays generally focus on the opposing side of the array, which is typically flat; see Fig. 3 (right). This flat side is known as the focal plane. It is at this plane that the micro-images are placed, one for every lens, side by side. Since each lenslet focuses to a point onto a micro-image below, an observer can never view two spots within a micro-image simultaneously; just one spot at a time, depending on the angle the observer looks though the lens. For example, if you have an array of small white dots, on an otherwise black background, behind each lens at the focal plane, any given lens will appear either completely black or white, depending on whether or not the lens is focused on a white dot, or the black background; see Fig. 3 (left). The state of each lens will vary depending on the point of observation. If all the dots are precisely ordered in a pre-calculated way, a completely different composite image can be directed to each eye of an observer, simultaneously, since each eye looks through the lens array at a different angle. The resolution of an integral image is therefore directly determined by the density of lenses in the array, since each lens effectively becomes a “dot”, or pixel (picture element), in the picture, with the visual state of each dot being a function of the viewing angle.
The first integral imaging method was “Integral Photography”. In this method the lens array is used to both record and play back a composite three-dimensional image. When an integral lens array sheet is brought into contact with a photographic emulsion at its focal plane, and an exposure is made of an illuminated object that is placed close to the lens side of the sheet, each individual lens (or pin-hole) will record its own unique micro-image of the object. The content of each micro-image changes slightly based on the position, or vantage point, of the lenslet on the array. In other words, the integral method produces a huge number of tiny, juxtaposed pictures behind the lens array onto the film. After development, the film is realigned with the lens sheet and a composite, spatial reconstruction of the object is re-created in front of the lens array, that can be viewed from arbitrary directions within a limited viewing angle.
Like lenticular, integral images can be created by digitally interlacing a set of pre-determined two-dimensional views to create three-dimensional and/or animation effects. Unlike lenticular, the imagery is viewable in all directions, within a limited viewing angle. Interlaced files can be printed using a variety of devices including ink-jet printers, film recorders, half-tone proofers, digital presses and press plates for lithographic reproduction. Unique, integral-specific, half-toning methods have been proposed by Dr. Daniel Lau and Trebor Smith; see Optics Express paper
While the mass-production of integral lens arrays remains limited, they are expected to become widely accessible in the near future as the relevant replication technologies continue to evolve. Once available, these lenses, when coupled with readily available digital interlacing and effects generation software, might enable lithographic integral imagery to develop as a significant advertising medium.