As NATO's 50th anniversary festivities in Washington draw to a close, a war continues to be waged on two fronts: one with the Serbian army in Yugoslavia, and on the other side, with the world press and public opinion. There appears to be little resistance on either front.

NATO feeds the press with daily briefings, including maps of targets, video clips of bombing runs, and occasional descriptions or aerial photography of alleged atrocities to bolster political resolve for escalation of the conflict. The press dutifully passes on the reports from NATO, with precious little means to verify them, and apparently very little effort to do so either. At best, the press has managed to indirectly corroborate some NATO stories with third-hand anecdotal evidence from fleeing refugees.

We all remember on Christmas Eve how the television weatherman would point out a superimposed image of Santa Claus flying across the radar screen. Our childhood naiveté allowed us the privilege of believing what we saw, and our parents had no trouble confirming to us that those images were real, and that we had better get to bed before Santa passed us by! Harmless propaganda, sure, but we were learning to defer to the experts.

When fighting a defensive war, it is easy to define the aggressor as the "bad guy," but when NATO reformed itself to become the offensive challenger, and policeman of Europe, it required a serious public relations campaign on the home front to convince its constituents that their cause was just. This propaganda may indeed be benign, and the cause may in fact be just. But any free and democratic society raining bombs on a sovereign nation must have a duty to constantly review the facts in detail to verify that it is on the "right" and "just" track.

The computers on our desks today are probably more powerful than what most CIA photo experts had 10 years ago at the end of the Cold War, so it seems like the public has a powerful set of tools to evaluate the propaganda from both sides. But just like in our childhood, we prefer to defer instead of analyzing with healthy skepticism what we are being shown on the 5 o'clock news.

Take, for example, the recent photo released by NATO as proof positive of refugee camps inside Kosovo.

The boxes in the left photo may indeed be "Makeshift Shelters" in displaced refugee camps, but without the labels at this resolution, one might also see it as a photo of milk cartons scattered in a park! This could be a Campgrounds of America near Atlanta, Georgia for all we know, but since NATO has placed an authoritative map nearby, we tend to trust their assessment.

Obviously, it would be unwise to distribute the full resolution images so that the enemy (and potential enemies) would be able to pinpoint our intelligence capabilities. But the question for civilian politicians supposedly in oversight of the military should be this, "If a lay-citizen cannot tell what a supporting photo shows without 'expert' labels, then is it ethical to use such a photo to justify political and military objectives to those citizens?" In other words, if one must rely on the experts for their opinion anyway, then any additional images for justification are just superfluous propaganda.

Imagine being able to convict someone for a crime based on the coarse pixelized images television uses to hide the identities of suspects on "Cops." For example, could you identify Monica Lewinsky from an 8x8 pixel array?

Another example of "content free" images can be found in the Izbica "Mass Graves" photographs.

The yellow NATO markings immediately indicate the expert assertion that this is a before and after picture of graves dug in a field. This begs for a simple, closer analysis.

The "after" picture (right) has very short shadows, so it can be assumed that this was taken near midday, making south be in the direction toward the bottom left of the picture. The "before" picture (left) has very long shadows to the lower right (toward the east), indicating that this photo must have been taken near sunset. There are no times or dates given by NATO for these photos, but if there really was some modification to the ground, then these would have had to have been taken on different days, which begs the first question: "Why did NATO take the 'before' picture at all?" There are a few farmhouses, but nothing of apparent strategic importance.

There are three possible explanations. (1) NATO somehow knew the area would be of importance later; (2) The pictures were not actually taken in the prescribed order, and the "sunset" photo was taken after the "midday" photo to check back on the odd formations; or, (3) NATO photographs every square foot of the theater, and recalls and compares all of the photo reconnaissance for activity. The latter is probably the most likely explanation, but that could be easily assessed by questioning the details of the photograph in a NATO press briefing, which has not been adequately attempted to date.

Another possibility is that the images were simply fabricated or modified. While there are many advanced methods for detecting such things (including walking up to the site with a news camera), a more simple (and safer) method is to simply look for irregularities in the color gradients of the image. Each pixel in a smoothed or blurred image is usually relatively close to the color of the next pixel. A false color filter can be used to highlight unusual color gradient artifacts, such as would be created with a "cut-n-paste" operation. Without much effort, a stair-stepped gradient around the added yellow marks becomes apparent in the image (probably from converting the image to JPEG after annotating).

It is also apparent that the "graves" have no obvious palette discontinuities, so they probably were on the original image, at least before the yellow lines were drawn on it in Belgium (or at the Pentagon).

Simply comparing known artifacts in the image to the unknown artifacts can reveal important information, such as comparing the "graves" to obvious things like trees, houses and shadows for size and color. A non-linear false-color transform helps highlight these comparisons. The image below sets the darkest one-percent of colors in the image to deep blue, and the brightest one-percent in the picture to red.

Surprisingly, blue can be seen on some of the "graves" (near midday remember.) These blue spots indicate a very dark direct shadow, exactly the same darkness as the big shadows of the house, the fences, and the bar ditches on the road. Therefore, the "graves" must either be high enough to cast a shadow, or deep as a well (and empty.) The nearby field shows the same dark blue, indicating perhaps some vegetative growth or other change from the "before" image.

The same dark color might be registered because the soil is that black, but then you would expect that color to be fairly uniform among the turned soil. The dark spots are not uniform, yet they are distributed like you might expect tall and short shadows from different people, for example.

Also, the orientation and placement of the "mass graves" near the farmhouses just doesn't make any sense. Why dig sideways up against a hill where there are big rocks? Why not dig in the soft tilled fields next to the road? And frankly, why would "mass graves" be dug individually instead of a single large pit such were discovered by Allied ground troops after World War II?

Considering what might be near a farmhouse, could the "graves" really be hay bales or hay stacks? Bales would explain the placement, and the shadows, but on closer inspection, the "graves" don't appear round or stacked ... they appear to be vertical. There is a distinct vertical rectilinear alignment of the rows of "graves." They are parallel with the walls of the houses and fences. They appear to be 5 or 6 feet tall (7-8 pixels tall, compared to 29 pixels for the 20-foot, two-story walls of the farmhouse). And the vertical alignment is not likely a resolution artifact caused by pixelization because the road and the fields have many odd-oriented holes and rocks that are smaller, and are obviously different in their alignment with the "graves."

Note the stick figures are drawn to scale to the houses for comparison with the "graves."

Whatever they are (or were), there appear to be three rows of about 25 dark vertical objects in a field at the bottom of a mountain. Are they Serb or KLA troops, haystacks, or a newly planted small orchard? Whatever they are, it should be difficult for any responsible citizen or journalist to accept these photos at face value as part of the rationale behind the war in Kosovo, even if it is justified on other grounds.