Tile Measurement

[Sherry Galeazzi]

The following tool estimates the total number of tiles necessary to cover a floor, roof, wall, or any other surface. It also considers the gap or overlap between tiles for better accuracy, as shown in the figure below.

Floor or wall tiles are typically installed with gaps between the tiles because the average tile may look quite similar to the next tile, but they are often not uniform in shape and size, and would not fit properly if installed without any gaps. These gaps are most commonly filled with grout, and as such, are often referred to as grout size, or grout lines. Grout is a form of concrete, and the gap between tiles can range from anywhere between one-sixteenth of an inch to half an inch in size. Different sized tiles, materials, and design needs all affect the size of the gaps. Although it is typically more difficult to have small gaps between tiles due to lack of a uniform shape and size, the use of rectified tiles (tiles that undergo additional processing to ensure that they are uniform) can allow for smaller spacing, though at an additional cost. For more uniformly cut tiles such as granite, smaller grout spacing can result in less visibility of grout lines between each tile.

In some cases, such as with roofing tile or the wood siding of a wall, rather than having a gap between them, tiles overlap to prevent leakage. The tile calculator can account for both of these situations. Either enter a positive value if there is a gap between the tiles being used, or a negative value if the tiles overlap.

Tile size can range anywhere from smaller mosaics that are 3/8", to 24" 48" slab tiles and everything in between. Square sizes (same width and length) are the most popular, accessible, and easiest to install. While straight edge tiles (rectangular, square, parallelogram) are the most common, unique tile shapes also exist, though installation is not as easy. Large tile sizes can make smaller rooms appear bigger, as well as more open and clean because there are fewer grout lines. However, installing larger tiles results in more wastage, while using smaller tiles can help add texture to a room.

There are a number of different classifications of tiles, including ceramic, porcelain, glass, quarry, and stone. Ceramic and porcelain tiles are the most cost efficient, and come in a variety of different styles. Glass tiles, while not appropriate for flooring because they crack under pressure, are visually unique and interesting; they are most commonly used for kitchen and bathroom backsplashes. Quarry tiles have rough surfaces that are good for floors that require grip, and are commonly used outdoors and in restaurant kitchens. Stone tiles include marble and granite, which provide unique and natural stone patterns, textures, and colors that are difficult to achieve using ceramics. They also offer the illusion of blending into grout edges, giving off an overall uniform look.

There are many different patterns used when installing tiles. The most common pattern used is a linear grid, with square or rectangular tiles, or a pattern involving angled squares or rectangles that form a typical diamond shape.

We make it easy to find out how many carpet tiles you need to order to create your perfect FLOR rug. Just select your preferred unit of measurement and enter the length and width of your rug (area will update automatically) to calculate the number of carpet tiles needed for your chosen area rug size.

Waste Factor will vary based upon tile size, layout, configuration of room, patterns, etc. Typical waste factor is about 10%. Add 15% for tile being installed diagonally or for a room with lots of jogs and corners. These installations will require more cuts and thus more waste.

Trim pieces and decoratives are typically sold by the piece. To figure the quantity you have to establish the length of the trim piece (i.e. 6" bullnose, 8" decorative liner), then the rule is: Linear Inches/Piece Length = Quantity

These formulas will help you estimate the quantities you'll need. We recommend you have a professional tile installer view the job to check for the suitability of your substrate and to measure and plan for any special conditions that may exist.

Does it 1) set the extension of the region from which all the tile features are computed, or 2) define a region used to compute a complementary set of features (the main tile features still being computed from the pixels lying inside the tiles).

If you want to use the ROI of any object, you select the ROI option. If you instead want a circle or a square at the object centroid (not the same as the center of thr bounding box) you select those options instead. Mostly used it for SLICs.

Ok, I see! Thank you for your answer. So, Tile diameter only applies when Square Tiles or Circular Tiles is used as Region (but not for SLIC Tiles or square Tiles). The coincidence is indeed a bit unfortunate and I would highly recommend to show Tile diameter in the dialog only when the former too are selected.

Those are two different things. Smoothing calculates averages of the values (well, not mean, gaussian based on distance I think?). The larger tile regions calculate the values across that region. For things like intensity means, this would likely be somewhat similar if the SLICs have similar areas. Though biased towards the center of all of the included objects. For the median, or Haralick features, it could be very different because you are suddenly taking a gaussian contribution from measurements that do not sum and divide (texture).

Rather, local measurements are scaled according to inter-centroid distances using scale values calculated from a Gaussian function. They are then added together, and re-normalized so that the sum of all the scale factors would add to 1. This avoids the number of nearby objects causing the measurements to differ substantially.

If you have multiple annotations (e.g. tile detections converted to annotations), and you want to append some custom measurement to the parent annotation relating to your detections (e.g. number of microglia per mm^2, percent area occupied by microglia), you can use the following script:

Alternatively, if you have experience in data visualization and statistical analysis in other programming platforms such as MATLAB, Python, or R, you could export the detection-level measurement and externally calculate your own metric for quantifying microglia abundance.

while also making sure to classify all of the detections at some point between adding the SLICs and adding the microglia. Although counting objects per SLIC sounds like a very suspect way of going about any sort of area or density measurement since SLICs vary in area, which invalidates using number of detections per tile as a measure of density.

Alternatively, if you are creating tiles anyway, create the tiles as annotations, which includes a count of how many detections are inside of them. Then classify the annotations based on detection object count. Would also require a script, however, to classify the annotations. And does not fix the underlying issue of the tile sizes being different, unless you were to use square tiles.

@Mark_Zaidi Thank you for your script, I think if I can figure out a bit more about how to resolve the hierarchy I can modify that to include the measurements in each tile. I like thresholding the density map, but I feel as though its a bit arbitrary

Regarding density mapping, im not sure how I would make it work. The rationale behind the tiling was to determine if cells are accumulating in each region. Microglia like to migrate to regions of pathology. So by classifying density in each tile I could figure out the percentage of tiles with say 3 microglia, 2 microglia and 1 microglia for 3+, 2+ and 1+ cell intensity classifications. Regions with high accumulation would be indicated by a higher proportion of 3+ tiles.

The main problem with measuring density by tile is that the boundaries are rather arbitrary. Somewhere in your image you have a 2 tile with a cell next to the border of a 0 tile- if you had shifted the borders of the rectangles just a little bit the cell would be assigned to the neighboring tile instead and you would have two 1 tiles instead of a 2 and a 0.

In contrast, density maps can look at the number of cells in the area around a pixel, and the boundaries of that measured region shift pixel-by-pixel so that there are no arbitrary cutoffs. The final results are much smoother and the objects created are more representative of the underlying biological phenomenon.

I have set-up some classes - high, medium and low density and thresholded accordingly. However I have an issue where the thresholded area overlaps into multiple anatomical structures(see yellow), how would one resolve this to get the density per region(black annotations)? The ultimate goal would be to have a percentage area of each region that is low density, medium density and high density

Hmm, apparently I was wrong about this. @petebankhead would it be difficult to include an option to add the density measurement based on a particular map to a given cell/object? Maybe I should open an issue for when you have time? There were some very old inefficient scripts that did this, but I have a feeling applying the value from the density map itself would be easiest.

For the moment you may want to try something more like: Cell Density Map - #18 by EP.Zindy
It will likely be slower, but it should add density values per class of object to the objects, which you can then use in an object classifier.

Im absolutely happy to keep looking into it though. As I do agree that the density maps are better for a whole slide approach, and if they could be segmented per region it would be far better than the tiles.

If you want to pursue this method, I can write some scripts to get you percent area of the yellow/orange/red regions in the brown annotations. But, it will be a few days until I can sit down and work on that. In the meantime, I think the tiling can work imperfectly but fine to get you the answers you need.

c80f0f1006