WO1995015063A1

WO1995015063A1 - Image depth improvement system

Info

Publication number: WO1995015063A1
Application number: PCT/AU1994/000732
Authority: WO
Inventors: Amedeo Filiberto Sala; Mickey Sala; Oscar Sala
Original assignee: Dr. Sala And Associates Pty. Ltd.
Priority date: 1993-11-26
Filing date: 1994-11-28
Publication date: 1995-06-01

Abstract

An image depth improvement system comprises a perforated mask (10) interposed at a prescribed distance between a viewer (12) and a 2-D monocular image on a display screen (14) of a TV monitor (16). The perforations in the mask (10) are adapted to introduce a perceived disparity in the visual information received by the left and right eyes of the viewer (12) so that an illusion of 3-D stereoscopic depth will be perceived by the viewer (12). The system may also include an image processing means (20) for conditioning selected regions of the image so that the perceived disparity introduced by the perforated mask (10) is enhanced. The image processing can be done at the receiver or at the point of broadcast. The system is fully compatible with conventional TV and computer display monitors.

Description

IMAGE DEPTH IMPROVEMENT SYSTEM

FIELD OF THE INVENTION

The present invention relates to an image depth improvement apparatus and method for producing an illusion of depth in an image on a display screen and relates particularly, though not exclusively, to such an apparatus and method for producing a stereoscopic 3-D effect in the 2-D monocular image on a screen of a conventional television receiver, computer display or bac lit display.

BACKGROUND TO THE INVENTION

Since the dawn of history, artists have attempted to reproduce three dimensional scenes by drawing, painting and statuary. As humans have the ability to see in three dimensions by virtue of stereopsis, they have always sought to create depth in art work. This was achieved in painting by inventing perspective, shading and "chiaro-scuro" , and by sculpting statues which provide full three-dimensional realism.

When trying to describe a series of related events - such as those on the Trajan column in Rome, as an example - artists invented bass-relief which, as the name implies, is a technique wherein complex scenes are depicted as a series of semi three-dimensional cartoons. Scenes sculpted in relief provide realism even if it can be viewed over an arc of only 180°, as is the case with TV wherein the small screens are more appropriate for bass-relief type images than full 3-D. The latter is more suitable for larger cinema screens.

The technology to create and display full 3-D images is complex and inherently expensive and its application suffers from a lack of programme material and would be incompatible with standard TV receivers. 3-D film and video are, in fact, useless unless the proper apparatus to form the illusion of stereoscopic images is provided to the viewer. The invention relates to the fact that humans favour three-dimensional images which artists try to simulate with shadows and perspective.

Prior art 3-D television and cinema systems have tended to concentrate on producing, transmitting, receiving and viewing binocular information to create a true stereoscopic or three-dimensional image. These systems involve reproducing the programme material using two cameras to produce a parallax effect; transmitting the stereo images to a receiver; decoding the stereo images at the TV receiver or cinema projector(s) and providing suitable viewing apparatus to enable a viewer to optically separate the two images viewed simultaneously or in rapid succession on the TV screen or cinema screen and to provide the two images to the left and right eyes respectively.

The present invention approaches the problem of creating an illusion of depth in the two-dimensional image on a display screen from a different angle altogether. It is based on certain psychophysical attributes of the human visual perception system that can be exploited to induce an illusion of 3-D stereoscopic depth in a monocular image. In the 1960's Prof. Bela Julesz of Bell Laboratories, N.J., USA did significant research into the human visual perception system. Throughout the description and claims, the term

"monocular image" refers to an image derived from a monocular source, ie. a single camera, as distinct from a binocular or stereoscopic image which is derived from twin sources, ie. two cameras giving slightly disparate views. The term "monocular information" refers to information received by the brain through one eye only.

The illusion of 3-D vision can be obtained by conveying - separately to each eye - images which are substantially the same. One is used as a reference image while the second carries details which are offset with respect to the reference image. This offset, also called parallax or disparity, is processed by the brain to create an illusion of depth.

Bela Julesz demonstrated that there is no need to have images with any meaningful monocular information to cause 3-D illusions. It is possible to have a strong and unambiguous impression of depth by means of random dot stereogra s wherein each image has no identifiable monocular information at all. Parallax is encoded by means of offsetting some of the random dots with respect to the reference stereogram. Depending on the direction of the offset, the image will either emerge or sink with respect to the surface upon which the stereograms are printed. This parallax is detected by the brain and translated into a 3-D image when the stereogram pair is looked at with the help of a stereo viewing apparatus or simply by crossing the eyes.

It has also been discovered that an impression of depth can be gained from parallaxes related to small regions of the images and therefore it is not necessary to present a true and complete set of stereograms to generate a 3-D illusion. When strong monocular information is present, the dependency on a "correct" parallax becomes much less important. This means that a parallax, that would give the illusion of a detail to sink in the Julesz experiment, can be made to emerge in the presence of strong monocular clues.

One reason for this phenomenon is that perspective and shadows help to complete the illusion of depth when only a few points of reference have been established by true or false offsets. It is not necessary to present both stereograms complete with all the stereo details, nor is it necessary to maintain precise information as to focus or dimensions in each stereogram.

While perspective and shadows are conveyers of the impression of depth, the sharpness of focus may be the most significant element in providing means of generating pseudo parallax details which can be used to create the illusion of 3-D. Generally a scene depends on this sharpness of focus to define a plane of reference. Anything out of focus is perceived as being either behind or in front of the subject presented in sharp focus. The true relationship of the relative position of the objects out of focus is derived from perspective and shadows. If details in sharp focus are presented separately to each eye, whereby an impression of parallax can be created, the image in focus will establish a plane of reference which will have all the elements to provide an illusion of three dimensional depth.

SUMMARY OF THE INVENTION

The present invention was developed with a view to providing an apparatus and method for image depth improvement which exploits certain psychophysical attributes of the human visual perception system to create an illusion of depth.

According to one aspect of the present invention there is provided an image depth improvement apparatus for producing an illusion of depth in a 2-D monocular image on a display screen, the system comprising: a display screen adaptor carrying a semi- transparent matrix interposed at a prescribed location between a viewer and the image, wherein said semi- transparent matrix is adapted to introduce a perceived disparity in the visual information received by the left and right eyes of the viewer whereby, in use, an illusion of depth will be perceived by the viewer.

Preferably said display screen adaptor is in the form of a perforated mask which includes a pattern of apertures, each aperture being of an optimum size selected relative to the minimum image element (pixel) size that the display screen is capable of resolving. Preferably the size and spacing of the transparent apertures is selected so that the semi-transparent matrix minimizes artefacts visible to the viewer. Two or more masks may be employed to introduce the perceived disparity. Typically the mask is positioned between 0 to lOcms in front of the display screen, more typically between 1 to 2cms in front of the display screen. However, the mask may be located at any suitable location between the viewer and the image, which results in a perceived disparity.

It may be possible in some images to produce a noticeable illusion of depth without any conditioning of the image. However, it is preferred that the video image be processed prior to display on the display screen in order to enhance the illusion of depth. For this reason, the system preferably further comprises: an image processing means for conditioning selected regions of the image wherein the perceived disparity introduced by the semi-transparent matrix is enhanced in the selected regions. Preferably said image processing means includes an attenuation processor for selectively attenuating the image in the horizontal and vertical directions so as to enable individual pixels to be distinguished from adjacent pixels. Preferably the image to be attenuated is sampled in a two-dimensional "checker board" array, however any suitable sampling pattern can be employed.

According to another aspect of the present invention there is provided an image depth improvement method for producing an illusion of depth in a 2-D monocular image on a display screen, the method comprising: introducing a perceived disparity in the visual information received by the left and right eyes of a viewer of the image whereby, in use, an illusion of depth will be perceived by the viewer.

Preferably, the perceived disparity is at least partially created by interposing a display screen adaptor carrying a semi-transparent matrix at a prescribed location between the viewer and the image.

Advantageously the method further comprises the step of: conditioning selected regions of the image so that the perceived disparity introduced by the semi- transparent matrix is enhanced in the selected regions. Preferably said step of conditioning involves selectively attenuating the image in the horizontal and vertical directions so as to enable individual pixels to be distinguished from adjacent pixels.

In a preferred embodiment of the present invention the perforated mask is placed in front of a display screen, and the displayed image is preferably modulated such that it has a regular pattern of similar spacing to the perforations in the mask, so that the viewer's left and right eyes see slightly "disparate" image detail. Binocular disparity of this type forms the basis of stereopsis, the process in which the brain can "fuse" the disparate left and right retinal images and estimate the depth of various objects in the 3-D world being perceived without any other depth cues.

The binocular disparity does not vary in the normal way according to the depth of the object imaged

(tending rather to be a fixed displacement which is least perceptible in brightly lit foreground areas) .

Nevertheless, the effect is sufficient to free the visual system from the conviction that the eyes are focused on a flat surface (the TV screen) and to suggest that a 3-D scene is being viewed. Released from the normal constraints imposed by evidence that the eyes are focused on a flat screen, the brain searches for the simplest interpretation of the incident light patterns (namely, those patterns arising from the original 3-D objects arranged in a 3-D world) , using whatever cues are available to determine relative 3-D positions. Significant advantages are that no special glasses are required, that the effect arises with conventional TV transmissions and that a wide range of viewing positions can be used. Furthermore, rotation of the viewing angle by tilting the head does not affect the 3-D effect. The 3-D effect achieved is quite unexpected and has a strong visual impact. Various cues contribute to the clarity of the 3-D effect, including: occlusion (objects in front obscuring objects behind them) ; relative movement (stationary objects in front moving across the image faster than objects behind as the camera moves) ; size of known objects; perspective; brightness; sharpness of focus; and shading/shadows. Cues which are normally available in perceiving a 3-D scene but not applicable here include binocular disparity, convergence and accommodation.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to facilitate a more complete understanding of the nature of the invention preferred embodiments of the image depth improvement apparatus and method will now be described in detail, by way of example only, with reference to the accompanying drawings in which: Figure 1 illustrates schematically a preferred embodiment of an image depth improvement system according to the invention; Figure 2 is an example of a magnified view of a

TV monitor screen used to illustrate how a perforated mask can introduce a perceived disparity;

Figure 3 illustrates an array of square pegs sticking out of a flat board and viewed monocularly (with one eye only) used to further illustrate how a perforated mask can introduce a perceived disparity;

Figures 4(a) and 4(b) illustrate interference patterns produced between the display screen and a perforated mask placed in front of the screen; Figures 5(a), (b) , (c) , (d) , (e) and (f) illustrate examples of possible mask patterns that may be used in the apparatus of Figure 1;

Figure 7(a) to (i) are timing diagrams for the block circuit diagram of Figure 6; Figure 6 is a block circuit diagram of one embodiment of an image attenuation processor that may be employed in the system of Figure 1;

Figures 8(a) and 8(b) illustrate a possible checker board sampling pattern employed by the attenuation processor of Figure 6 for interlaced and non-interlaced display screens respectively; and,

Figure 9 illustrates graphically a attenuation function employed by the attenuation processor of Figure 6.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

A preferred embodiment of the image depth improvement system according to the invention for producing an illusion of depth in a video image on a display screen is illustrated schematically in Figure 1. The system comprises a display screen adaptor 10 carrying a semi- transparent matrix interposed at a prescribed distance between a viewer 12 and the image on a display screen 14 (not visible), in this case, of a TV monitor 16. The semi- transparent matrix of the adaptor 10 is adapted to introduce a perceived disparity in the visual information received by the left and right eyes of the viewer 12 so that an illusion of depth will be perceived by the viewer 12. The adaptor 10 is typically placed at a prescribed distance from the surface of the screen 14 upon which an image will be reproduced. This distance may be established empirically by supplying a random noise signal to the TV monitor 16 instead of a video signal. The ideal distance is that where a strong illusion of three-dimensional randomly floating dots is perceived by the viewer. If the viewer is located within normal viewing range, ie. half a metre or more, from the screen 14, the adaptor 10 would typically be located between 0cm to 10cm in front of the display screen 14. Usually the adaptor 10 is located approximately 1cm to 2cm in front of the display screen and may, for example, be attached to the glass front face of the TV monitor 16. The image depth improvement system illustrated in

Figure 1 also includes image processing means 20 for conditioning selected regions of the image so that the perceived disparity introduced by the semi-transparent matrix of the mask 10 is enhanced in the selected regions of the image. In this embodiment, the image processing means 20 comprises an attenuation processor for selectively attenuating the image in the horizontal and vertical directions so as to separate individual pixels from adjacent pixels. The apparatus and method for effecting image attenuation will be described in greater detail below.

The semi-transparent matrix of adaptor 10 may be formed by a perforated opaque plastic or acetate material having a pattern of dots or lines provided therein. In some applications, for safety reasons, the perforated mask may be made from metal in order to absorb screen radiation. The size of the transparent holes or apertures of the semi- transparent matrix is optimally selected relative to the minimum image element (pixel) size that the display screen 14 is capable of resolving. Furthermore, the size and spacing of the transparent apertures is selected so that the semi-transparent matrix is ideally suited to minimize artefacts visible to the viewer 12. The apertures may also be selected with an appropriate geometric form that correlates with the attenuation function. Examples of typical perforated mask patterns are illustrated in Figures 5(a) to (f) .

The semi-transparent matrix of the display screen adaptor may take the form of a liquid crystal (L.C.) array in which the aperture size and two dimensional geometry can be made programmable. A variable three dimensional geometry of the mask may also be useful in controlling the perceived disparity. Furthermore, small lenses could be used in the semi-transparent matrix in place of apertures.

An important advantage of the image depth improvement system of this embodiment is that no internal modifications to the TV monitor 16 or to the television transmission signal are required. In other words, the image depth improvement system is fully compatible with conventional television transmission facilities and TV receivers and video monitors. As shown in Figure 1, the television or video signal from a conventional VCR tuner player 22, a video conferencing codec 24, computer graphics/video games 26, a laser video disc player 28 or other hardware or software source is processed in the processing means 20 before being supplied to the TV monitor 16. The processing means 20 may be in the form of a "black box" which sits next to the TV monitor 16 in the viewer's home, or alternatively the processing means 20 may be located at the point of broadcast (TV transmitter/service provider) .

Whilst it is preferred to include some form of electronic image conditioning in order to enhance the illusion of depth, it is still possible to produce a noticeable illusion of depth using one or more perforated masks without any conditioning of the image. For example, in one embodiment of the invention (not illustrated) two masks are employed to create the disparity in the visual information received by the left and ^•right eyes of the viewer as described above. One mask is adhered to the front face of the TV monitor, whilst the other is separated from the display screen by a small distance, for example, 1 or 2cms. When the image is viewed through the two masks, an interference pattern is observed with parallax between the left and right eye. The mask adhered to the front face of the monitor performs a attenuation function by reducing the intensity of the light passing through it. The combination of the two masks creates the interference pattern and disparity to achieve a 3-D effect. The perforated masks may have identical patterns of transparent holes for one type of effect, or they may be different to achieve another type of effect. In either case, a degree of depth of separation can be achieved.

In an alternative embodiment, a single perforated mask is used with a hole size which is sufficiently small to match the RED-GREEN-BLUE phosphor dots on the picture tube screen of the TV monitor. The interaction between the perforated mask and the tube may achieve the necessary left eye and right eye disparity (parallax) to create the same 3-D effect. Figure 2 represents an example of a magnified view of a TV monitor display screen 14. Each group of three squares 30 correspond to the RED-GREEN-BLUE phosphor dots of the television picture tube screen. If a sufficiently small resolution perforated mask is placed in front of the screen, an interference pattern is observed. The interference pattern is obtained as a beat interference between the TV screen phosphor dots and the perforated mask. The interference pattern regulates the intensity of light that passes through the perforated mask. The disparity introduced by this interference is caused by the varying light intensity of individual pixels as perceived by the left and right eyes of the viewer. This disparity works in conjunction with normal 2-D cues in the image to produce the 3-D effect observed. Figures 4(a) and (b) illustrate the interference pattern produced between the TV display screen 14 and a perforated mask 10 when viewed monocularly in different positions. In the figures, the resolution of the mask is approximately 2.5 times less than the resolution of the display screen. Even though the interference patterns are quite distinct, their effect when viewed with two eyes (in stereo) is to introduce a disparity (parallax) in the image as perceived by the left and right eyes. This interaction of the mask with the picture image is perceived to dramatically induce a 3-D effect. The 3-D effect cannot be observed in the two dimensional drawings on paper as they are an illustration of the interference effect observed when viewed with one eye only.

The following discussion outlines an explanation of why the interference patterns observed through the perforated mask are present, and how it is thought they induce a 3-D effect. If a small 1cm sticker, say black, is placed in the centre of a blank A4 sheet of paper, and then another sticker, say green, is placed on top of the black sticker with a small offset of a few millimetres, one would perceive the green sticker as having depth on a two- dimensional plane as though it were a cylinder appearing to be projecting into the page. The effect is similar to an artist's pencil sketch that uses shadowing to give perspective or depth to a drawing. If the green dot is moved a few millimetres in any direction, then a changing apparent depth and angular projection will be observed.

If many black and green dots are placed symmetrically over the entire page, all having the same displacement with respect to each other, the page would then appear to have many cylinders projecting in one direction (exhibiting depth) . Furthermore, if the cylinders are replaced by wooden pegs sticking out of a flat board, and viewed with one eye looking in the middle of the pegs, one would see only the square end of the middle peg and the rest of the pegs appearing to gradually lean outwards towards the sides of the board. Figure 3 illustrates one vertical column and one horizontal row of square pegs protruding from a flat board and viewed monocularly looking at the middle peg.

If one were to imagine the square ends of the pegs, or perhaps better, the green stickers, to be the holes of a perforated mask placed 2cms in front of another identical mask and viewed with one eye, one would observe the same effect as the pegs or stickers, with the centre hole appearing substantially aligned with the rear mask, and the neighbouring holes progressively misaligning and appearing as small cylinders leaning outwards. An important difference with the perforated mask is that the holes are relatively small, and the angular viewing over the area of the entire mask allows repetitive sequences of leaning cylinders going outwards and then slowly inwards in a sine function to be seen, repeated sequentially in both the vertical and horizontal directions. This is what causes the observed interference patterns visible, for example, in Figures 4(a) and (b) .

The observed effect arises due to the shadowing caused by misaligned viewing angles of the perforated mask holes and those of the rear mask, or in the case of a attenuated image, the attenuated area (and/or picture pixel cylinder) misalignments. In actual pictures, image modulated, pixellated cylinders are in fact at work to produce a sense of depth. When viewed with both eyes, image cylinders are leaning in one direction through one eye, and in a slightly different direction through the other. When the brain fuses these two disparate images, an illusion of 3-D depth is perceived. Added to this are learned and conditioned perspective interpretations that the brain uses, combined with super-edging and filtering effects, so that a 2-D image can be translated into 3-D. To reduce the visibility of the interference lines, the perforations in the mask are preferably made as fine as possible, and the image attenuation is preferably formed with a resolution to maximise picture quality and the 3-D effect.

As noted above, it is preferred that the video image be conditioned prior to display on the display screen in order to optimise the illusion of depth. Preferably the image is processed to ensure that disparity is extracted only from that part of a scene which must appear as the 3-D reference plane. Conditioning of selected regions of the image can be performed using spatial to frequency domain transform processing, such as continuous or discrete Fourier Transform (DFT) techniques or equivalent algorithms which enable details of the image to be manipulated. The algorithm defines areas of high resolution by utilising transform coding of localised spatial 2-D pixel correlation. These parameters are typically monitored at all times before processing the image to identify the selected regions to be attenuated.

The processing of images in the processing means 20 of Figure 1 may typically involve any one or more of the following steps:

(i) examining the frequency components of the image to define which object is in sharpest focus. (ii) making a decision as to whether the object should be further blurred or focused.

(iii) distinguishing between small isolated details and large objects and determining by the degree of overlapping, during motion, their relative positions. (iv) increasing the focus sharpness of the edges of the image which define the reference planes.

(v) attenuating the image in the horizontal and vertical directions to optimise left and right eye image separation. (vi) prefiltering or image enchancement.

If the TV image is in sharp focus, foliage and vegetation will appear to have a distinct three-dimensional "body" . The image processing typically ensures that fine details are maintained only in the zone of interest thus avoiding ambiguous 3-D effects. The above method can also improve image fidelity. A sharp rise in virtual resolution is experienced. This helps to engage the visual system into the "super-edging" mode whereby edges are perceived as being sharper. This is because the human visual system "connects" the points which are relevant to the object, but discards points which are perceived as not belonging to it. This helps to filter out artefacts such as overshoots and noise peculiar to analog signal handling.

The image processing means 20 of the preferred embodiment includes an attenuation processor similar in some respects to the DOT depletion grid described in copending international application No. PCT/AU93/00368. A preferred embodiment of the attenuation processor employed in the image depth improvement system of Figure 1 is illustrated in block diagram form in Figure 6.

Figure 6 illustrates a preferred embodiment of the attenuation processor 40 for sampling the incoming video signal and attenuating selected regions of the image according to an attenuation function described below. The analog input video signal at 42 passes through an analog filter 44 before it is converted to an eight bit digital format by the analog to digital converter (ADC) 46. The function of the analog filter 44 is to perform as a passband filter providing high frequency emphasis (increasing sharpness) of the image. The digitized video signal from ADC 46 then passes through a digital filter 48 before it is attenuated according to the attenuation function in attenuation circuit 50. The function of the digital filter 48 is to perform any one or more of steps (i) to (v) and (vi) above. The attenuated digital output of the attenuation circuit 50 is sampled by a 2 to 1 multiplexer (MUX) gate 54 under the control of the logical A/B switch control 52. The MUX 54 samples the A/B switch control 52 so that either the attenuated digital signal or the original unattenuated digital video signal from the digital filter 48 is allowed to pass through the MUX 54 (but not both simultaneously) . The digital output of MUX 54 is then converted to an analog signal by the digital to analog converter (DAC) 56 to produce an analog output video signal at 58 which can be fed directly to the video input of a television receiver. The attenuation processor 40 also includes a timing circuit 60 for generating a timing signal supplied to the MUX gate 54 for sampling the attenuated digital video signal from the attenuation circuit 50. . The operation of the timing circuit 60 will now be described with reference to the timing diagrams illustrated in Figures 7 (a) to 7 (i) .

The horizontal synchronisation signal is separated from the analog input video signal by horizontal sync separator 62 to produce a horizontal sync signal illustrated in Figure 7 (a) . The horizontal sync signal is then supplied to a first monostable circuit 64 which produces a delayed output signal illustrated in Figure 7 (b) . The delayed output signal of the first monostable circuit 64 is then supplied to a second monostable circuit 66 which produces an output signal as shown in Figure 7(c) . A pixel clock generator 68 produces a square wave signal having a frequency corresponding to the horizontal pixel sampling frequency, as illustrated in Figure 7 (d) . For simplicity, the illustrated pixel clock signal only provides nine sampling periods per horizontal scanning line. The frequency of pixel clock generator 68 shown in Figure 6 is to be distinguished from the true pixel data rate which is used by the analog to digital converter (not shown in Figure 6) where the true pixel data rate may be higher to meet the full pixel resolution of the source video signal. It should be understood that the pixel clock frequency shown in Figure 6 may be the same as the true pixel data rate or it may be at a lower data rate so that the terms "attenuated pixel" and "unattenuated pixel" used throughout this description may cover more than one true pixel in the source video signal.

The pixel clock generator 68 ultimately defines the duration of the A/B switch control 52 on the MUX 54 and thereby the number of true consecutive pixels attenuated horizontally. Unlike the true pixel data rate which is fixed, the pixel clock in Figure 6 can be varied to match the spatial number of the attenuated true pixels with the desired display screen adaptor aperture size thereby minimising the interference artefacts . While not shown in figure 6 the same consecutive number of true pixel attenuation/non-attenuation variability can be achieved in the vertical direction.

The attenuation function used in this description does not alter the true pixel resolution of the video signal. It selectively attenuates the brightness or intensity of individual true pixels.

The output signal of pixel clock generator 68 is divided by two at block 70 to produce a signal as shown in Figure 7(e) . This signal is then supplied to an exclusive OR (XOR) gate 72. The output signal of the second monostable circuit 66 is also divided by two at block 74 to produce a signal shown in Figure 7(f) . This signal is also supplied to the XOR gate 72 which therefore produces an output signal as shown in Figure 7(g) . The output signal of the second monostable circuit 66 is also inverted by inverter 76 to produce an inverted signal as shown in Figure 7 (h) , and this signal together with the output signal from the XOR gate 72 is supplied to a NOR gate 78. The output signal of NOR gate 78 is a modified square wave as shown in Figure 7(i), in which the square wave of alternate horizontal scanning lines is 180° out of phase with the previous scanning line to produce a checker board sampling pattern as shown in Figure 8(a) . The output of logical NOR gate 78 is then digitally buffered by the logical A/B switch control gate 52. The buffered output of A/B switch control gate 52 is then sampled by MUX 54 A/B input line to control the output of MUX 54. The block circuit diagram illustrated in Figure 6 shows the attenuation process applied in a digital format, however substantially the same process can also be carried out using analog circuitry. An analog realisation of the attenuation processor 40 could employ the same timing circuit 60, however the ADC 46 and digital filter 48 would be replaced by an analog buffer. Likewise, the MUX 54 and DAC 56 would be replaced by a video amplifier for attenuating the original analog video signal to produce a conditioned output video signal. Furthermore, substantially the same attenuation process may be embodied in appropriate computer software. For example, a full frame of the video image is stored in a video graphics memory of a PC, and selected pixels of the stored image are then attenuated according to the attenuation function and simultaneously displayed on the VGA monitor. The attenuated video image stored in the graphics memory may be further processed and then converted from a VGA format to an output video signal for supply to a television receiver. In one application of this embodiment of the attenuation processor, the pixel brightness (luminance) is attenuated (reduced in intensity) according to a attenuation algorithm described below. The attenuation processor samples the picture image as a two dimensional checker board array as illustrated in Figures 8 (a) and 8(b) . The pixels corresponding to the white squares on a checker board are left unaltered, whereas the adjacent pixels corresponding to the black and shaded squares are attenuated (reduced in intensity) as described below. For illustration purposes only, the diagrams in Figures 8(a) and 8(b) depict low resolution frames of 13 x 10 pixels. In reality, the horizontal and vertical resolutions are much higher than shown. For example, 740 x 548 for conventional TV and 1024 x 768 for computer display monitors. The picture attenuation resolution is preferably selected to match the resolution of the perforated mask used in the system. It has been found that a noticeable 3- D effect can still be achieved with as low as 1/8 normal TV viewing resolutions.

Throughout the description and claims the term "conditioning" should be understood in a general sense to include any processing of the individual picture elements (pixels) of a video image which results in a modification of the pixels in selected regions of the image in such a way that the perceived disparity in the image is enhanced to improve the illusion of depth. Conditioning the image pixels may involve attenuation of the luminance or chrominance components or both, and may also include more complex signal processing.

Figure 8(a) illustrates an interlaced display used in television, wherein two fields are required to make up one frame. The odd and even lines in the checker board array correspond to the odd and even fields respectively of one frame. Shaded and black squares in Figure 8(a) correspond to pixels which are conditioned by the attenuation function over alternate fields in the interlaced mode. Black squares in Figure 8(b) correspond to pixels which are conditioned according to the attenuation algorithm over each frame in the uninterlaced mode. Furthermore, Figure 8(b) can also represent one field of an interlaced frame, in other words the same sampling pattern is used in odd and even fields. The 3-D effect works in both interlaced and uninterlaced modes.

Figure 9 illustrates graphically the input versus output function of a typical attenuation algorithm employed by the attenuation processor of Figure 6. In the illustrated embodiment, the video signal is sampled with an analog to digital conversion (ADC) producing eight bits or 256 grey scales. For a given input pixel intensity, the output pixel intensity is attenuated. The shaded region in Figure 9 shows the difference between the input and output pixel values according to the attenuation algorithm. The attenuation algorithm may be linear or non-linear. The digital output is then converted to an analog video signal with a digital to analog converter (DAC) . The analog video signal can then be input directly to a television monitor.

As is evident from Figure 9, the degree of attenuation is a function of the brightness of the input pixels. For example: a very bright (white) pixel with grey scale of

255 decimal, would undergo no attenuation; a white grey pixel with grey scale of 200 would undergo only moderate attenuation; a grey pixel with grey scale of 128 would undergo more than moderate attenuation; a grey black pixel with grey scale of 80 would undergo significant attenuation; and, a very dark pixel with grey scale of 16 or less would be completely attenuated to black. In most images, darker pixels are correlated to backgrounds. Attenuation biased towards darker areas of the image creates an interference pattern with the perforated mask producing disparity in these areas . Depth is then perceived more easily in the backgrounds or more dramatically where light to dark changes in picture areas occur. Limiting attenuation in the light areas of an image increases luminance in those areas, which in turn increases the general contrast. It also minimises disparity in these areas. These areas of the image become a reference plane that differs from adjacent attenuated areas which appear at a different depth. Where image plane differences are perceived, an image depth enhancement is produced.

In the above embodiment, the attenuation algorithm is "keyed" to brightness intensity (luminance) . The algorithm may be keyed to colour components (chrominance) RED GREEN or BLUE hues or pallete. For example, the composite video signal may be converted to the respective RED GREEN BLUE colour components and then conditioned individually. The conditioned colour components are then input directly to a suitable monitor, or alternatively reconstituted back into the composite analog signal for display on a normal TV monitor. More intelligent algorithms may use motion detection to decide which parts of the image to condition. For example, still areas of the image could use the above attenuation function, whereas motion areas may use a different function.

Now that a preferred embodiment of the apparatus and method of image depth improvement have been described in detail, it will be apparent that the system has significant advantages over conventional 3-D systems, including (but not limited) to the following:

(a) the system employs a relatively inexpensive display screen adaptor (perforated mask) and conditioning hardware; v

(b) a distinct 3-D stereoscopic effect can be achieved in a 2-D monocular image without the use of special 3-D glasses or other viewing apparatus; (c) there is no requirement for special cameras to shoot video/film in "stereo";

(d) the system can be easily and inexpensively incorporated into conventional TV receivers; (e) the system hardware can be incorporated either at the point of broadcast transmission or at the receiving home TV set.

Numerous variations and modifications will suggest themselves to persons skilled in the electronics and visual perception arts, in addition to those already described, without departing from the basic inventive concepts. For example, the system of image depth improvement can be modified for use in connection with large screen display systems and/or backlit advertising displays of the kind employed in the fast food industry. Some of these displays use dithered print to produce colour (RED, GREEN and BLUE colour hues are produced by halftone density dithering (or offset printing) the primary colours in the display) . Often, soft drink vending machines use this form of display. A perforated mask placed about 1cm from the sign can achieve an illusion of depth. If a dithered printing technique is used on the display, which may roughly approximate the output of the attenuation processor, then a 3-D effect may be observed. All such variations and modifications are to be considered within the scope of the present, the nature of which is to be determined from the foregoing description and the appended claims.

Claims

THE CLAIMS DEFINING THE INVENTION ARE AS FOLLOWS:

1. An image depth improvement apparatus for producing an illusion of depth in a 2-D monocular image on a display screen, the system comprising: a display screen adaptor carrying a semi- transparent matrix interposed at a prescribed location between a viewer and the image, wherein said semi- transparent matrix is adapted to introduce a perceived disparity in the visual information received by the left and right eyes of the viewer whereby, in use, an illusion of depth will be perceived by the viewer.

2. An image depth improvement apparatus as defined in claim 1, wherein said display screen adaptor is in the form of a perforated mask which includes a pattern of apertures, each aperture being of an optimum size selected relative to the minimum image element (pixel) size that the display screen is capable of resolving.

3. An image depth improvement apparatus as defined in claim 2, wherein the size and spacing of the apertures is selected so that the semi-transparent matrix minimizes artefacts visible to the viewer.

4. An image depth improvement apparatus as defined in claim 3, wherein the mask is positioned between 0 to lOcms in front of the display screen.

5. An image depth improvement apparatus as defined in any one of claims 2 to 4 further comprising: an image processing means for conditioning selected regions of the image wherein the perceived disparity introduced by the semi-transparent matrix is enhanced in the selected regions.

6. An image depth improvement apparatus as defined in claim 5, wherein said image processing means includes a attenuation processor for selectively attenuating the image in the horizontal and vertical directions so as to enable individual pixels to be distinguished from adjacent pixels as perceived by the viewer.

7. An image depth improvement apparatus as defined in claim 5, wherein said image processing means includes means for sampling the image according to a pattern which is correlated with the pattern of apertures in the perforated mask.

8. An image depth improvement apparatus as defined in claim 7, wherein said sampling means samples the image according to a checker board pattern.

9. An image depth improvement apparatus as defined in claim 6, wherein said attenuation processor includes an attenuating circuit for attenuating selected pixels in the image according to an attenuation function, wherein the degree of attenuation is a function of the brightness of the pixels.

10. An image depth improvement method for producing an illusion of depth in a 2-D monocular image on a display screen, the method comprising: introducing a perceived disparity in the visual information received by the left and right eyes of a viewer of the image whereby, in use, an illusion of depth will be perceived by the viewer.

11. An image depth improvement method as defined in claim 7, wherein the perceived disparity is at least partially created by interposing a display screen adaptor carrying a semi-transparent matrix at a prescribed location between the viewer and the image.

12. An image depth improvement method as defined in claim 8, further comprising the step of: conditioning selected regions of the image so that the perceived disparity introduced by the semi- transparent matrix is enhanced in the selected regions.

13. An image depth improvement method as defined in claim 9, wherein said step of conditioning involves selectively attenuating the image in the horizontal and vertical directions so as to enable individual pixels to be distinguished from adjacent pixels as perceived by the viewer.

14. An image depth improvement method as defined in claim 12, wherein the step of conditioning selected regions of the image involves sampling the image according to a pattern which is correlated with a pattern of apertures provided in the semi-transparent matrix.

15. An image depth improvement method as defined in claim 14, wherein the image is sampled according to a checker board pattern.

16. An image depth improvement method as defined in claim 12, wherein said step of conditioning selected regions of the image involves attenuating selected pixels in the image according to an attenuation function wherein the degree of attenuation is a function of the brightness of the pixels.