3D Image Enhancement made EASY – Part III.

Comparing Simple Contrast Methods – cont'd.

Part I (April 25th.) of this series introduced a new method of image enhancement (simple '3D') and gave examples of its use.

Part II (May 3rd.) offered further examples of results using this method, particularly as compared to Brightfield and COL.

Part III continues this effort with examples at higher magnification and higher resolution.

Now, in casual microscopy, considerable time is often spent using the 40x ("high dry") objective. This lens offers the combined advantages of reasonably high magnification and resolution, decent working distance, and ease-of use. Also, almost every standard microscope has one. Further, more than 90% of what maybe seen (at high magnification) with Transmitted Light Microscopy may be seen using this type of lens. So, it seems only proper that we should determine just how well the 3D Mask method works with this lens…

In the first set, below, we compare results obtained using Brightfield with those using the Radial 3D mask method. This  set shows the same diatom photographed using each of these methods, with the results adjusted to approximate the actual visual appearance of the object. (If anything, these images understate the advantages of the 3D method, in part due to loss of image quality due to the use of the JPEG image format during image processing.)

Click anywhere on the above image for larger versions. 

Note that the measured width of diatom used in these photo is approximately 21 microns, which results in a calculated 'dot spacing' of about 1.0 microns. As this is reasonably within the resolution capability of the NA 0.70 objective, the performance of the objective itself should not be a limiting factor in these tests.

Not apparent in the photo set is the slight loss in overall image brightness associated with the 3D mask. However, this loss is a small penalty to pay for increased image contrast and resolution.

Radial 3D versus 'COL' – more surprises? 

In the second set, below, we compare the Radial 3D mask method with COL, a currently popular alternative method of contrast enhancement. Once again, the relative levels of image brightness are not preserved in these images, but here, COL typically suffers a much greater loss of overall brightness (typ. 3 to 4 times greater) than does the 3D mask method.

Click anywhere on the above image for larger versions. 

Here the 3D method appears to be at a disadvantage, at least initially, but only because the Condenser Iris was fully open (e.g: 100%).

However, as shown in the third photo set (below), this limitation is easily overcome by simply reducing the Iris opening.  This results in a sort of "variable contrast" mode, where the overall image contrast is readily controllable by means of the Condenser Iris. For Iris openings down to about 60%, there is very little loss in resolution, while even smaller openings (down to about 40-50%) may be used to achieve further increases in image contrast.

Click anywhere on the above image set for larger versions. 

Note that the COL method has no equivalent mode!

With COL results are fixed, as determined by the annulus diameter and opening width. To adjust contrast with COL either the entire Iris must be exchanged or a variable condenser, such as the scarce (and costly) Leitz 'Heine' Condenser, must be used. (Note that the Heine condenser will typically fit only a few older microscopes, whereas the Radial 3D mask method may be used on nearly any microscope equipped with a standard Condenser!)

Finally, remember that, with COL, usually the entire condenser annulus (or the Heine condenser setup) must be changed with every change of objective, while, with the Radial 3D mask method, only a simple Iris "tweak" is all that is typically be needed, if at all.

Next, in Part IV of this series, we will explore some of the unique possibilities for even greater image enhancement using the Radial 3D mask method…


* * * * *

3D Image Enhancement made EASY – Part II.

Comparing Simple Contrast Methods

A recent post (April 25, below) has revealed a very simple method of optical contrast enhancement, "Radial 3D Effect", which also adds a measure of 3D-like effect to an overall contrast improvement.. The method is very easy to implement and to use, functioning with a wide range of common objectives and yet requiring no special adjustments with objective changes.

Yet, in spite of these these and other operational advantages (described below), the basic question remains, "Just how well does this new method work?"

In this post we begin to address that question, starting with a direct comparison of images from this new method with those from two common alternative methods: (a) Simple Brightfield, and (b) Circular Oblique Lighting (or, COL). (These other methods also share the traits of simplicity, ease-of-use, and of not requiring any special objectives.)

So, how do these methods actually compare? 

To see, we can begin by examining the image set below… 

Click anywhere on the above image set for larger versions. 

The top image is in Brightfield mode, with illumination by the Kohler Method, with the Condenser Iris set to about 100% of the full objective aperture. This is to serve as the "Reference" image. It reveals a reasonable amount of detail but one could hope for greater contrast. Also, the image appears "flat," revealing little of the object surface texture.

The middle image is basically identical, but with the addition of the Radial "3D" Effect mask discussed earlier. Note that this image shows not only increased object contrast and detail, but also reveals object surface contours, especially evident in the specimen on the left.

The final image was made by replacing the "3D" mask with a standard Condenser phase ring (annulus) to produce COL contrast. Note that here there is also increased contrast (relative to Brightfield), but the "3D" effect seen in the middle image is absent.  However, much unlike the Radial 3D Mask method, image results with COL can be quite dependent upon the exact choice of annulus, as well as 'annulus-specimen' matching. Thus, with COL, each objective potentially requires using a different annulus. (The 'Radial 3D Mask' technique has no such sensitivities.)

One important factor, not shown by the above images, is the difference in overall image brightness which occurs with each of these methods.

Brightfield, of course, offers the highest image brightness level, with the Radial '3D' Mask method running second with somewhat less than half that level (depending on the exact diffusion material used). The COL method, however, is far behind this with overall brightness at 15% or lower, depending upon the size and width of the Condenser annulus selected. (Note that some specially-made COL rings may support slightly higher overall image brightness levels.)

Condenser iris Effects

Both Brightfield and Radial 3D methods also support a basic level of user control over "depth-of-focus" in the specimen plane, simply by adjusting the Condenser Iris opening. Reducing the opening, even by a small amount (e.g: from 100% to 80%) can result in a significant increase in not only depth-of-focus, but also in overall image contrast. These effects are depicted in the image set  below:

Click anywhere on the above image set for larger versions. 

As expected, reducing the Condenser Iris opening in either Brightfield or 3D mode provides a slight increase in image contrast, but in the 3D mode this more significantly enhances the apparent depth-of-focus for the specimen. In general, Iris openings from 100 percent open, down to 50 or 60 percent open, can prove useful. At less than about 50% open, image resolution can begin to suffer noticeably.  

Unfortunately, the COL method (described above) does not directly support such control – with COL the only recourse is to switch to using a smaller diameter annulus, which often can have unpredictable and undesirable effects on the overall appearance of the specimen. 

Be aware that the image set presented above is not intended as definitive basis for comparing these methods, but merely as a basic indication of the results produced by each method. More precise imaging techniques for these three methods, as well as for some additional methods, will be needed to permit extending this evaluation to a more definitive level. 

Note that the above images are limited by minor variations in focus between modes, as well as a slight residual camera motion which may obscure finer detail. Both these issues will be addressed in the test setup before any additional comparison photos are produced.

This comparison will continue is a subsequent post, or two…

* * * * * *