TEM Image Analysis Platform

ChemFitX provides an online TEM image analysis workflow for researchers working with transmission electron microscopy images of nanoparticles, nanostructures, catalysts, powders, thin materials and related specimens. The platform combines calibrated manual measurements, automated particle-size and morphology analysis, grayscale intensity analysis, statistical summaries, size-distribution plots, image annotation and export tools in one browser-based workspace.

Upload a TEM image, calibrate the image from a visible scale bar or a known pixel size, and then choose between manual measurements and automated particle analysis. The automated workflow supports region-of-interest selection, multiple thresholding methods, bright- or dark-particle polarity, optional morphological cleanup, touching-particle separation and particle-by-particle morphology measurements.

The TEM image is read and analyzed locally in your web browser by this application. The uploaded microscopy file is not sent to ChemFitX servers during the analysis.

What the TEM Image Analysis Platform Can Do

Import TEM images in TIFF/TIF, PNG, JPG/JPEG and BMP formats.
Calibrate image scale from a visible scale bar or from a validated direct pixel-size value.
Measure distances, diameters, radii, areas, perimeters, widths, heights and angles using interactive drawing tools.
Use line, circle, ellipse, rectangle, triangle, polygon, angle, point, arrow and cross tools for measurement or annotation.
Calculate manual measurement statistics and generate a size-distribution histogram from the primary linear dimensions of drawn measurement objects.
Analyze the original decoded image as 8-bit grayscale and report mean, median, standard deviation, minimum, maximum, RMS contrast and Shannon entropy.
Display a grayscale-intensity histogram for the whole image or a selected rectangular region of interest.
Perform automated particle segmentation using Otsu, adaptive local-mean or manual thresholding.
Analyze either bright particles on a darker background or dark particles on a brighter background.
Filter segmented objects by minimum and maximum pixel area.
Apply optional morphological opening, closing, or opening plus closing to clean the binary mask.
Use conservative connected-object analysis or watershed-style distance separation for touching or agglomerated particles.
Optionally exclude particles that touch the analysis boundary.
Report particle count, equivalent circular diameter, major and minor axes, aspect ratio, circularity and accepted area fraction.
Provide particle-level area, perimeter, orientation, centroid coordinates and edge-touch information in exports.
Plot particle distributions by equivalent circular diameter, major axis, minor axis, area, circularity or aspect ratio.
Overlay detected particles and optional particle IDs on the TEM image for visual verification.
Adjust display brightness, contrast and threshold without silently replacing the original decoded pixels used for quantitative analysis.
Pan, zoom, rotate, flip, center and inspect the image interactively, with undo and redo support.
Customize measurement-line color, transparency, width and solid/dashed/dotted line style.
Export an annotated PNG image and numerical results to Excel, CSV or JSON, with dedicated particle CSV and Excel exports.

Supported TEM Images and Analysis Requirements

Supported image formats

The platform accepts TIFF/TIF, PNG, JPG/JPEG and BMP images. For quantitative microscopy work, use the highest-quality original image available. Avoid unnecessary resizing, resampling or repeated JPEG compression because these operations can alter object boundaries, apparent dimensions and intensity information.

A valid spatial calibration

Measurements can be reported in physical units only after the image has been calibrated. ChemFitX supports scale-bar calibration and direct unit-per-pixel calibration. If no calibration is set, measurements remain in pixels.

Representative image quality

Automated particle analysis works best when particles have sufficient contrast from the background and when the selected image region is representative of the sample. Strong overlap, low contrast, contamination, scale bars, labels, image borders and unrelated structures can interfere with segmentation and should be excluded with a region of interest when appropriate.

Scale Calibration for TEM Measurements

Scale-bar calibration

Select the calibration tool and draw a line across the full visible scale bar. Enter the known scale length and choose its unit. The platform supports nm, μm, mm, cm and Å for scale-bar calibration. ChemFitX then determines the pixel-to-length conversion used by manual and particle measurements.

Direct pixel calibration

If the physical pixel size is already known from the microscope acquisition or validated metadata, enter that value directly. Direct calibration supports nm/pixel, μm/pixel, Å/pixel and mm/pixel.

Why calibration accuracy matters

Every physical size, area and perimeter calculated from the image depends on the spatial calibration. A scale-bar line that does not match the full printed bar, an incorrect unit, or an image that has been resized after the scale bar was generated will propagate directly into the reported dimensions.

Manual TEM Measurement Tools

Line and arrow measurements

Line and arrow tools report calibrated length. They can be used for particle dimensions, spacing between features, layer thickness, pore dimensions or other direct distances that are clearly resolved in the image.

Circle measurements

A circle measurement reports radius, diameter, area and circumference. This is useful for approximately circular particles or features when a circular geometric approximation is scientifically appropriate.

Ellipse measurements

An ellipse measurement reports the two ellipse radii together with area and an approximate ellipse perimeter. It can be useful for elongated or anisotropic features that are reasonably represented by an ellipse.

Rectangle, triangle and polygon measurements

Rectangles report width, height, area and perimeter. Triangles report the three side lengths, area and perimeter. Polygons report area, perimeter and the number of vertices. These tools provide geometric measurements for irregular regions that are better represented by manually defined boundaries than by a simple circle or ellipse.

Angle, point and cross tools

The angle tool reports the angle between two segments. Point and cross tools are primarily visual markers that can be used to identify positions or features of interest on the image.

Manual Measurement Statistics and Size Distribution

ChemFitX automatically summarizes the primary linear dimensions of manually drawn measurement objects. The manual statistics panel reports the number of included objects, mean size, standard deviation, minimum size and maximum size, and it generates a size-distribution histogram.

These manual statistics are intentionally separate from the automated particle population statistics. They summarize the selected primary linear dimensions from drawn objects and should not be interpreted as a complete automated particle-size analysis unless the measurements were collected using a consistent and representative sampling procedure.

Automated TEM Particle Size and Morphology Analysis

The automated particle module segments image regions into particle objects and calculates population-level and particle-level descriptors. Before running the analysis, choose whether the entire image or a rectangular region of interest should be analyzed.

Region of interest

A rectangular ROI can be drawn to exclude scale bars, text labels, blank margins, damaged regions or unrelated image areas. When enabled, the same ROI can be used for both automated particle analysis and grayscale intensity analysis.

Thresholding methods

Three segmentation modes are available. Otsu thresholding automatically selects a global threshold from the grayscale distribution. Adaptive local-mean thresholding calculates a local threshold using a user-defined window and offset, which can be useful when illumination or background intensity varies across the image. Manual thresholding gives direct control over the 0–255 segmentation level.

Particle polarity

Choose bright particles on a darker background or dark particles on a brighter background. Selecting the correct polarity is essential because the threshold mask determines which pixels are treated as particle regions.

Area filtering and morphological cleanup

Minimum and maximum pixel-area filters can remove objects outside the desired size range. Optional morphological opening can remove small isolated specks, while closing can fill small gaps in segmented objects. Opening plus closing applies both operations. These controls modify the segmentation mask and should be used conservatively.

Touching-particle separation

The conservative connected-object mode treats each connected segmented region as one object. A watershed-style distance-separation option is also available for touching or agglomerated particles. Watershed separation can improve object separation in suitable images, but it can also over-segment noisy or irregular particles, so the overlay should always be checked visually.

Boundary exclusion

Particles touching the analysis boundary can be excluded. This can reduce bias from partially visible objects whose complete size and shape are not contained within the analyzed region.

Particle Morphology Results

Equivalent circular diameter

Equivalent circular diameter (ECD) is the diameter of a circle having the same measured area as the segmented particle. It provides a convenient single-number particle-size descriptor for objects with different shapes, but it does not imply that the particles are actually circular.

Major and minor axes

The platform reports moment-based major- and minor-axis estimates. These values summarize the directional extent of an object using an ellipse-like representation and are useful for comparing elongated versus more equiaxed particles.

Aspect ratio

Aspect ratio describes particle elongation from the major- and minor-axis relationship. Values closer to unity indicate a more equiaxed ellipse-like shape, while larger values indicate greater elongation.

Circularity

Circularity is calculated as 4πA/P², where A is particle area and P is particle perimeter. A value near 1 indicates a compact shape close to a circle, while lower values can result from elongation, irregular boundaries, rough edges or segmentation noise.

Area fraction

Accepted area fraction reports the percentage of the analyzed region occupied by the accepted segmented particle objects. It is an image-area metric and should not automatically be interpreted as a bulk volume fraction, mass fraction or concentration.

Additional particle-level data

Particle exports can also include area, perimeter, orientation, centroid coordinates and whether an object touched the analysis boundary. These descriptors can support further statistical analysis outside the browser.

Particle Size and Morphology Distributions

The automated particle histogram can display distributions for equivalent circular diameter, major axis, minor axis, area, circularity or aspect ratio. This makes it possible to inspect both particle size and shape heterogeneity rather than relying on a single mean value.

Distribution quality depends on segmentation quality and image sampling. A large number of segmented objects does not by itself guarantee a representative population if the TEM field of view is biased toward a particular region or if overlapping particles are systematically merged or split.

Grayscale Intensity and Contrast Analysis

The intensity module analyzes the original decoded image as 8-bit grayscale values from 0 to 255. Grayscale is calculated from the decoded RGB values using the weighted relation 0.299R + 0.587G + 0.114B.

Intensity statistics

The platform reports mean grayscale, median grayscale, standard deviation, minimum and maximum grayscale values, RMS contrast and Shannon entropy. RMS contrast is calculated as grayscale standard deviation divided by mean grayscale intensity. Entropy is calculated from the 8-bit grayscale histogram and summarizes the diversity of intensity values in the analyzed region.

Intensity histogram

A grayscale histogram shows the distribution of pixel intensities across the whole image or selected ROI. It can help compare contrast distributions between images acquired and processed under comparable conditions.

Higher-bit-depth TIFF images

If a higher-bit-depth TIFF is loaded, the platform can report the source bit depth when available, but the browser intensity analysis uses the decoded 8-bit representation. Therefore, this module should not replace dedicated quantitative high-dynamic-range microscopy analysis when the original high-bit-depth pixel values are required.

Image Display, Navigation and Annotation

The workspace includes select/edit, pan, zoom, reset zoom, rotation, horizontal flip, vertical flip, image centering, fullscreen viewing, undo and redo. The zoom control supports detailed inspection without changing the underlying physical calibration.

Measurement lines can be customized by color, transparency, line width and solid, dashed or dotted style. Annotation text can also be customized, and exported images can optionally include measurement labels with user-selected font size, family and weight.

How to Use the ChemFitX TEM Image Analysis Tool

Upload a TEM image in TIFF/TIF, PNG, JPG/JPEG or BMP format.
Inspect the image and identify a reliable scale bar or confirm that a validated physical pixel size is available.
Calibrate the image using the scale-bar tool or direct pixel calibration.
For manual analysis, choose a measurement tool and draw the required line or shape on the feature of interest.
Repeat manual measurements when a statistically meaningful size summary is required, and review the manual statistics and histogram.
For automated particle analysis, draw an ROI if scale bars, labels, blank margins or unrelated regions should be excluded.
Choose Otsu, adaptive local-mean or manual thresholding and select the correct particle polarity.
Set minimum and maximum particle area filters and apply morphological cleanup only when it improves the segmentation mask.
Choose conservative connected-object analysis or watershed-style separation for touching particles.
Run particle analysis and verify the segmentation overlay before accepting the numerical results.
Review particle count, ECD statistics, area fraction, circularity, aspect ratio, the particle table and the selected distribution plot.
Use the optional intensity module when grayscale statistics and contrast descriptors are useful for the analysis.
Export the annotated image and numerical results to the required format.

Interpreting the Main TEM Results

Manual size statistics

Manual mean, standard deviation, minimum and maximum summarize the primary linear dimensions extracted from user-drawn objects. They are only as representative as the manual sampling strategy used to create them.

Particle count

Particle count is the number of segmented objects accepted after thresholding, area filtering, optional morphology, separation and edge-exclusion rules. Changing segmentation settings can change the count substantially.

Mean, median and SD of ECD

Mean ECD describes the arithmetic average particle size, median ECD gives the central value of the ordered size population, and the standard deviation describes the spread of ECD values. For skewed particle-size distributions, the median and histogram can be as important as the mean.

Circularity and aspect ratio

Circularity and aspect ratio describe different aspects of morphology. Circularity is sensitive to perimeter irregularity as well as elongation, while aspect ratio primarily reflects the relationship between the estimated major and minor axes.

Intensity metrics

Grayscale mean, RMS contrast and entropy describe the decoded image intensity distribution. They should be compared across samples only when imaging, detector, exposure, processing and scaling conditions are sufficiently comparable for the intended scientific interpretation.

Exporting TEM Analysis Results

Annotated PNG image

The image export saves the TEM image with optional measurement overlays. Export options allow measurement labels to be included or omitted and allow the label font size, font family and font weight to be adjusted.

Excel

The main Excel export contains manual measurement data and can also include particle and intensity sheets when those analyses have been performed. The dedicated particle Excel export includes particle-level data together with a summary of particle count, accepted area fraction, segmentation settings, analysis region and scale unit.

CSV

The main CSV export provides manual measurement records. A separate particle CSV export includes particle area, equivalent circular diameter, major axis, minor axis, perimeter, aspect ratio, circularity, orientation, centroid coordinates and edge-touch information.

JSON

JSON export stores the manual measurement results in a structured format suitable for reuse or programmatic processing.

Data Privacy and Browser-Based Processing

For this TEM platform, the uploaded microscopy image is read and processed locally in the browser. The analysis code does not upload the TEM image to ChemFitX servers.

Users should still follow their laboratory or institution policies for unpublished microscopy images, confidential research data, shared computers, browser storage and downloaded exports.

Scientific Scope and Limitations

The platform is an image-analysis aid and does not replace microscope calibration, acquisition-quality control or expert TEM interpretation.
Physical dimensions require a valid scale calibration. Incorrect scale-bar measurement, unit selection or pixel-size input produces correspondingly incorrect physical measurements.
Manual measurements depend on user placement of lines and boundaries and should be repeated when measurement uncertainty or heterogeneity is important.
Automated particle results depend strongly on thresholding, image contrast, particle polarity, ROI selection, area filters and morphological settings.
Touching particles may be merged in connected-object mode, while watershed-style separation can over-segment noisy or irregular objects.
Particle boundaries obtained from thresholded TEM contrast are image-segmentation boundaries and may not correspond perfectly to the true physical particle boundary.
Equivalent circular diameter is an area-equivalent size descriptor and does not imply that particles are circular.
Moment-based major and minor axes are ellipse-like shape estimates and should not be interpreted as exact crystallographic dimensions.
Circularity is sensitive to segmentation noise and perimeter roughness as well as real particle morphology.
Area fraction measured in a 2D TEM image is not automatically equivalent to bulk volume fraction, mass fraction or number concentration.
Objects touching the image or ROI boundary may be incomplete; excluding them can reduce truncation bias but also changes the sampled population.
Display brightness, contrast and threshold controls are visualization aids; quantitative particle and intensity analysis use the original decoded image pixels together with their own analysis settings.
The intensity module analyzes an 8-bit decoded representation. Higher-bit-depth source data are not preserved as high-bit-depth numerical intensity values in this browser analysis.
Grayscale intensity metrics are affected by microscope settings, detector response, exposure, image processing and specimen thickness and should not be compared across unrelated acquisition conditions without justification.
JPEG compression can alter particle edges and intensity values; use original TIFF or other lossless source images when available.
A single TEM field of view may not represent the full sample. Population conclusions require an appropriate sampling strategy across sufficient images and regions.

Frequently Asked Questions

What is TEM image analysis?

TEM image analysis uses transmission electron microscopy images to quantify features such as particle size, shape, area, spacing, geometry and grayscale intensity. Reliable interpretation requires appropriate calibration, image quality and sampling.

Can I measure particle size online with ChemFitX?

Yes. You can measure particles manually with calibrated drawing tools or use automated particle segmentation to calculate population-level size and morphology descriptors.

Which TEM image formats are supported?

The platform accepts TIFF/TIF, PNG, JPG/JPEG and BMP images.

Are my TEM images uploaded to ChemFitX?

For this application, the image is read and analyzed locally in your browser and is not uploaded to ChemFitX servers by the analysis code.

Do I need a scale bar?

A visible scale bar is one way to calibrate the image. If a validated physical pixel size is already known, you can use direct pixel calibration instead.

Which automatic thresholding methods are available?

The particle module supports Otsu global thresholding, adaptive local-mean thresholding and manual threshold selection.

Can ChemFitX analyze both bright and dark particles?

Yes. The particle-polarity setting supports bright particles on darker backgrounds and dark particles on brighter backgrounds.

Can the tool separate touching nanoparticles?

The platform includes a watershed-style distance-separation option in addition to conservative connected-object analysis. The segmentation overlay should be checked because touching-particle separation can under- or over-segment objects.

What is equivalent circular diameter?

ECD is the diameter of a circle with the same area as the segmented particle. It is a convenient size descriptor for irregular shapes but does not mean the particle is circular.

What does circularity mean?

Circularity is calculated as 4πA/P². Values closer to 1 correspond to more compact circle-like segmented shapes, while lower values can reflect elongation, irregularity or noisy boundaries.

What does aspect ratio mean?

Aspect ratio describes elongation using the estimated major and minor axes. Larger values indicate a more elongated object.

Can I analyze only part of an image?

Yes. You can draw a rectangular ROI and use it for particle and intensity analysis.

Does changing display brightness or contrast change the numerical particle results?

The display adjustments are for visualization. Automated particle analysis uses the original decoded image pixels together with the dedicated segmentation settings.

What intensity parameters are calculated?

The platform reports mean and median grayscale, standard deviation, minimum and maximum intensity, RMS contrast, Shannon entropy and an 8-bit grayscale histogram.

Can I use a 16-bit TIFF for quantitative intensity analysis?

The image can be loaded and the source bit depth may be reported, but this browser intensity analysis uses the decoded 8-bit representation. Use dedicated high-bit-depth workflows when original high-dynamic-range values are required.

Can I export particle-level data?

Yes. Dedicated particle CSV and Excel exports include individual particle measurements, and the main Excel export can also include particle and intensity information when available.

Can I export an annotated TEM image?

Yes. The platform can export a PNG with optional measurement labels and configurable export text styling.

Scientific References

Williams, D. B.; Carter, C. B. Transmission Electron Microscopy: A Textbook for Materials Science. Springer.
Gonzalez, R. C.; Woods, R. E. Digital Image Processing. Pearson.
Otsu, N. A Threshold Selection Method from Gray-Level Histograms. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9, 62-66.

Vincent, L.; Soille, P. Watersheds in Digital Spaces: An Efficient Algorithm Based on Immersion Simulations. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1991, 13, 583-598