During the year 2000, the MGS spacecraft acquired two partial images of the Mars Face, M1400709 and M1600184. The second image contained the most information and will be discussed first.

Eye-Like Feature in MGS Image M1600184

On June 6, 2000, the MGS high-resolution camera acquired a new image of the Face landform at Cydonia, image number M1600184. The first MGS image, SP122003, was acquired on April 5, 1998.

While the newer image shows only a partial view of the already-photographed western side of the Face, it is superior to the first image for the following reasons:

• The resolution was 1.54 meters, which is almost three times greater than the resolution of the first image, with a resolution of 4.3 meters.
• The viewing conditions were evidently better, without the atmospheric haze that partially obscured the landform in the first image.
• The first image was taken at a camera viewing angle 45-degrees from the local vertical. The off-nadir line of sight caused features in the horizontal plane of the planet surface to be foreshortened, which effectively reduced the resolution of the camera in the east-west direction to about 70% of the resolution in the north-south direction. For the second image, the MGS spacecraft was almost directly over the landform so there was no foreshortening of features in the horizontal and therefore no significant loss of resolution in any direction.

A comparison of the two images shown in Figure 1 indicates that the grotesque appearance of the Face in the first image was in fact largely attributable to the extreme lighting conditions, exacerbated perhaps by the haze and off-nadir viewing angle. The first image shows a vaguely almond-shaped dark patch suggestive of a glaring, rather poorly drawn eye. The same feature in the second image exhibits a complex structure far more similar to a human eye. It consists of a bright object within a well-defined almond-shaped depression corresponding to the iris and sclera (white) of a human eye. Further, this structure resides in what appears to be a larger but shallower depression, also almond-shaped, that corresponds to the hollow of the eye socket between the forehead and cheekbone of a human face.

Figure 1. Left: Face from MGS image SP122003. Right: Portion of Face in MGS image M1600184 around eye-like feature.

The labeled arrows indicate features in the images as follows:

1. Iris-like convex object.
2. Dark, probably concave, area corresponding to the eye socket of a face.
3. Boundary of well-defined concavity corresponding to the lower edge of an eyelid.
4. Inner apex of C.
5. Outer apex of C.

The complexity of this area of the Face in image M16000184 is more clearly apparent in the larger-scale version of Figure 2. The

Figure 2. Close-up of "eye" region of Face in M1600184.
 
 

Figure 2 shows the "eye" region with no digital magnification of the original image. At this scale, the two inner apexes of the "eye" structure can be seen quite clearly. A large portion of the upper edge of the "eye" concavity also appears to be well defined. There is an outer apex (farthest from the center of the Face) on the rim of the concavity that would be expected for a human eye, although not as distinct as the two inner apexes.

It can be seen that the "iris" is roughly triangular with a somewhat faceted appearance and not circular like the iris of a human eye. This could perhaps be explained as a the application of a technique sometimes used by sculptors to produce a glint of light within a sculpted eye that makes it more easily seen from a distance.

Also of interest is the cellular texture of the "eye-socket" region surrounding the "eye," which seemingly radiates out from the central concavity. This texture, which is strongly suggestive of some artificial modification of the landform, could be another sculpting trick. The texturing of this area could make it darker than the smoother surrounding regions. This texture difference could give an observer the impression that the concavity is much deeper than it really is, thus allowing the creators to avoid the necessity of the massive excavation necessary to produce a depression with a depth in true proportion to the dimensions of a face.
 
 

Confirmation of DiPietro's Predictions

Years before the Mars Global Surveyor was launched, SPSR member Vince DiPietro subjected the Viking images of the Face to a false-color enhancement that highlighted all pixels within a selected range of gray scale values. The resulting enhancements showed a circular "eyeball" area within the shadowed "eye socket" on the western side of the Face. Due to the low resolution of the Viking Orbiter camera, the reality of this feature has been disputed. The measurements described here show that the eye-like feature in M1600184 is in fact near the position predicted by DiPietro.

To determine the position of the "eye" in the MGS image on the Viking image, both were resized so that the distance from the right edge of the "teardrop" formation to the left edge of the mesa was the same in both. The MGS image was adjusted to account for the aspect ratio of 1.16 and skew angle of 93 degrees given in the ancillary data for the image. Measurements of points for comparison of the Viking and MGS image were made relative to an X-Y coordinate system defined as follows:

The straight left edge of the landform was taken as the Y-axis. The X-axis was taken as a line tangent to the curved top edge of the "headdress" and perpendicular to the Y-axis. The origin (0, 0) point of the system is then the intersection of these two lines.

The (x, y) coordinates of the "iris" in the MGS image are (85, 120) in this coordinate system as shown in Figure 3. A point at the same coordinates in the Viking image is at the left edge of the blue-shaded area that DiPietro termed the "eyeball." It seems reasonable to conclude that the MGS "iris" and "sclera" structures correspond to the "eyeball" in DiPietroís enhancement of the Viking image, therefore satisfying the prediction of an eye-like feature.

Figure 3. Plot of the position of the "iris" and "teardrop" features in the second MGS image M1600184 onto DiPietroís false-color enhancement of Viking Frame 35A72.
 
 

Estimate of Measurement Uncertainty

The Viking image in Figure 3 used to plot the position of the "iris" was enlarged by a factor of 8, so the uncertainty in the position of any point is about 8 pixels, or +/- 4 pixels. The uncertainty of the relative position of the center of the "iris" and the right edge of the "teardrop" should then be on the order of 8 pixels. Taking into consideration other sources of uncertainty, such as the differences in lighting between the two images, it seems reasonable to estimate the uncertainty to be on the order of 10 pixels. The size of this 10-pixel radius of uncertainty is shown in Figure 3.

The horizontal distance from the right-hand side of the "teardrop" was used for scaling the two images. Note that the vertical coordinate of that point in the MGS image falls reasonably close to the correct vertical distance in the Viking image from the top of the headdress to the "teardrop." This gives some grounds to believe that the coordinate system used is as accurate as the uncertainties permit.
 
 

Image M140079: Part of the "Mysterious" Eastern Edge of the Face Revealed

In April 2000, two months prior to the acquisition of the image showing the eye-like structure on the western side of the Face landform, MGS acquired an image that shows a very small part of the eastern side of the Face. This is the same side that was in shadow in the Viking images, greatly foreshortened in the first MGS Face image, and not included at all in the June, 2000 MGS image. This earlier image, M1400709, although very limited in the area covered, does provide one piece of significant information that is worth noting. It confirms that even the best enhancements of the first MGS Face image, SP122003 from April 1998, give a highly misleading impression of the shape of the eastern side of the landform.

Figure 5 below shows the section of M1400709 that includes a small part of the lower-right (eastern) side of the Face. In the center, the complete landform from SP122003, the first MGS image, is shown at about 1/3 full size. On the far right, the eastern edge of the Face from SP122003 is shown at full size. The points A and B identify the same positions in all three images. Point A is the position of a bright feature just beyond the eastern edge of the Face. This feature was used as a point of registry. Point B is the position where the eastern edge of the Face intersects the edge of M1400709.

Figure 4. Left: Well-defined eastern edge of Face from M1400709, full scale; Center: Face from SP122003 1/3 full scale; Right: Apparently featureless eastern edge of Face from SP122003, full scale. Letters A and B indicate corresponding points in the three pictures.

Note: the center image of Figure 4 showing the entire Face landform is only partially rectified. While it correctly shows the two-dimensional shape of the base perimeter, it distorts the 3-dimensional shape, most critically, it puts the position of the "nose" ridge much farther from the vertical centerline of the Face than it really is.
 
 

It can be seen that even at full scale, SP122003 suggests the eastern edge of the landform is poorly defined and shows little detail. It seems to be a featureless surface that slumps down gradually to the surrounding plains. In contrast, the western edge of the landform is well defined in SP122003. Some have argued that this lack of definition of the eastern side is evidence that the Face is really an irregularly shaped natural landform. Others have argued that the apparent slumping of the eastern side could be due to erosion of what was once a more symmetrical shape.

Image M1400709 shows that both arguments are irrelevant, at least for the length of the eastern edge of the Face that it includes. Even though M1400709 is of significantly lower resolution than SP122003 (6.18 meters/pixel versus 4.38/pixel), it shows that the eastern edge is very well defined, and may even be close to a mirror image of the western edge. Resolution is not the only factor that determines the quality of an image. The lack of definition of the eastern edge and probably much of the rest of the landform in SP122003 can be completely attributed to the off-nadir viewing angle, the hazy atmospheric conditions at the time it was taken, camera exposure and gain settings, or some combination of those factors. The intrinsic short-comings in the raw image were, of course, magnified and compounded by the improper techniques that JPL used to process the first "enhancement" they released to the news media in 1998.