DCM File Opener & DICOM Viewer — Open .dcm Online, No Upload

Q: How do I open a .dcm file?

Drop it on the box at the top of this page. A .dcm file is a DICOM file - the format every hospital scanner writes - and this page is a DCM file opener: it reads the file in your browser and shows the scan, with nothing to install and nothing uploaded. It works the same on Windows, on a Mac, on Linux, on a Chromebook and on a phone, because the only program involved is the browser you already have. If what you were given is a disc or a folder rather than one file, drop the whole folder. A CT or an MR is one file per slice, often with no .dcm on the end of the names at all, and opening them together is what puts the scan back in order. The DICOMDIR file and the Windows viewer that came on the disc can be left where they are; neither is needed.

Open a DCM file

Drop a scan here or click to browse — one .dcm file, or a whole folder of them

Drop a whole folder's worth of files at once and they are put back into their series and stacked in order. Nothing is uploaded, and nothing is written back to your files.

{name}: {why} {what} ({files}) {a} · {b} ⁦{width} × {height}⁩ ⁦{width} × {height}⁩, {frames} {n} {unit} {n} compressed fragment, {bytes} bytes {n} compressed fragments, {bytes} bytes {bytes} bytes, not shown (empty) {bytes} bytes: {head} {bytes} bytes: {head} … {n} item {n} items January February March April May June July August September October November December {day} {month} {year} {date}, {time} {n} day {n} days {n} week {n} weeks {n} month {n} months {n} year {n} years MONOCHROME1 (0 is white) MONOCHROME2 (0 is black) PALETTE COLOR (indexes into the lookup tables) YBR_FULL (colour, as luminance and two differences) YBR_FULL_422 (colour, with the differences at half width) HFS (head first, supine) HFP (head first, prone) FFS (feet first, supine) FFP (feet first, prone) M (male) F (female) O (other) {size} ({bytes}) ⁦{row} × {column}⁩ mm (none) (private) (unknown) there is no DICM marker at byte 128, and the first bytes are not a DICOM element either this fragment does not start with a JPEG SOI marker a JPEG scan arrived before its frame header this fragment is a DCT-based JPEG, not a lossless one this fragment carries no JPEG frame header a JPEG frame with no size this frame subsamples its components, which lossless JPEG here does not support a JPEG frame of {bits} bits, which is outside the 2 to 16 the format allows this scan names a Huffman table the file never defined a Huffman code longer than the sixteen bits the format allows the RLE segment header is 64 bytes and this fragment is shorter an RLE frame declares {count} segments, and the format allows 1 to 15 this image needs {wanted} RLE segments and the frame has {count} RLE segment {segment} points outside the fragment this file declares no image size this file has no pixel data fragments {syntax} is a compression this page cannot decode frame {frame} is past the end of the pixel data no pixel data fragment holds frame {frame} {bits} bits per sample, which is not 1, 8, 16 or 32 this file has {fragments} pixel data fragments for {frames} frames and no offset table to say which belong to which the file ends at {available} bytes; {wanted} more were needed at {at} This file has no DICM marker, so it was read as a bare dataset in {syntax}. That was worked out from its first element rather than declared. The file meta group names no transfer syntax, so the dataset was read as Implicit VR Little Endian, which is what the standard says to assume. Transfer syntax {uid} is not one this tool recognises. The header below was read as Explicit VR Little Endian, which every syntax registered since 1995 uses; the pixels could not be decoded. The file ends part-way through the element at byte {at}; everything before it was read. The element {tag} declares an undefined length, which only a sequence may do. Nothing after it could be read. The element at byte {at} says its value is {length} bytes, and only {left} are left in the file. DICOM header — {name} File size {count} bytes Transfer syntax Object Image {bits}-bit greyscale {count} samples Frames Pixel spacing Notes on reading this file File meta information Dataset Read in a browser at {origin}, which uploads nothing. Drag across the picture to widen or narrow the window, and up or down to move its centre. Both numbers are below, and typing into them does the same thing. Drag to move the picture inside its frame. The scroll wheel zooms, and “Fit” puts it back. Drag a line across the picture to measure it. Where the file says how far apart its pixels are, the answer is in millimetres; where it does not, the answer is in pixels and says so. {value} at column {column}, row {row} {length} px — this file does not say how far apart its pixels are, so this cannot be in millimetres Stacked by where each slice sits in the patient, which is the order the scanner took them in. Stacked by instance number: these files carry no position in the patient to sort on. The gap between slices is not the same all the way through, so at least one slice is missing from this series. This file’s pixels are compressed with {codec}, which no browser can decode and which this page does not carry a decoder for. Everything else about the file is below. The picture could not be decoded: {reason}. The header below was read in full. This file carries no pixel data at all. It is a header, and everything in it is below. These pixels are baseline JPEG, so the browser’s own decoder unpacked them and what it hands back is eight bits deep. The window still works; the precision the scanner recorded is not all there. {shown} of {total} elements. {hidden} more. All {total} elements in this file. {total} elements match “{query}”. One element matches “{query}”. Nothing matches “{query}”. Show the remaining {count} private element not in this dictionary The file did not say; this is what the data dictionary gives for this tag. Reading {name}… Reading {at} of {total}… Nothing here could be read as DICOM. {why} This tool reads the DICOM format itself, so it has nothing to say about other files. One file was skipped: {files} {count} files were skipped: {files} Series {number} 1 file {count} files Slices are {gap} apart. {at} of {total} your scan has gone nowhere. {total} files loaded, all of them this page's own. {platform} something contacted {hosts}, which this tool never does. Treat that as worth investigating. {platform} The page's own ad, measurement and donate-button scripts loaded from {hosts} host; not one of them was given a scan, a pixel of one, or a word from its header. The page's own ad, measurement and donate-button scripts loaded from {hosts} hosts; not one of them was given a scan, a pixel of one, or a word from its header. frame {at} of {total} instance {number} From the file From the file, {n} Everything in this frame Set by hand Soft tissue Lung Bone Brain Liver Mediastinum Angiography {count} frames no pixel data not stated in this file signed unsigned {bits} bits, {sign}, 1 sample per pixel, {photometric} {bits} bits, {sign}, {samples} samples per pixel, {photometric} {low} to {high} colour — nothing measured to report Nothing in this file names or numbers a person. The identifiers a scan usually carries are either absent or empty. It also carries {count} identifiers of its own — the study, series and instance UIDs. None of them is a name, and every one of them is a perfect key into the archive that made the file. And {count} private elements, whose meaning is published nowhere: some scanners keep a second copy of the patient’s name in one. Copied. Your browser would not let the page copy. The download button beside it writes the same text to a file. Saved {name}.

View as Markdown

# DICOM Viewer — a DCM file opener that runs in your browser

CT, MR, X-ray and ultrasound, with the window, the header and the measurements.

> A free DCM file opener: open .dcm scans — CT, MR, X-ray and ultrasound — in your browser, with nothing to install. Window and level, scroll a whole series, measure in millimetres, read every DICOM tag, and see exactly what in the file identifies the patient. Nothing is uploaded.

This is an interactive browser tool at https://abox.tools/dicom-viewer/. Files are processed on the device running that browser. Any deliberate sharing or optional network feature is described on the page. What follows is its written explanation; open the address to use it.

## Your scans are **processed in your browser**. No processing upload.

The scan is opened and decoded by your own browser: the header, the pixels, the window, the measurements. There is no server on the other end of this page to send protected health information to even if anything here wanted to, and nothing about the file - not the patient's name, not the study, not the filename - is read out to anyone.

- ✗ No upload
- ✗ No account
- ✓ Works offline
- ✓ Open source
- ✓ Files stay on your device

## How to open a DCM (DICOM) file

1. **Choose the files.** One `.dcm` file, or the whole folder off the disc — a CT or an MR is one file per slice, and dropping all of them at once is what puts the series back together. Files are read straight off your disk by the browser; nothing is sent anywhere while you do it.
2. **Pick the series.** A study usually holds several: the scout, then each acquisition. Each one is stacked in the order the scanner took it in, worked out from where each slice sits in the patient rather than from the numbering, which is not always in the same direction.
3. **Set the window.** This is the control that makes a scan readable, and the one an image editor does not have. Drag across the picture to widen the window and up or down to move its centre, or pick one of the named windows — lung, bone, brain, soft tissue — on a CT, where the units are the same on every scanner in the world.
4. **Scroll the stack.** The slider under the picture moves through the slices, and the arrow keys do the same thing once you have clicked on the picture. A multi-frame file — an ultrasound loop, an angiogram — plays with the button beside it.
5. **Measure something.** Switch to Measure and drag a line. Where the file says how far apart its pixels are, the answer is in millimetres and takes account of pixels that are not square; where the file does not say, the answer is in pixels and says so rather than inventing a scale.
6. **Read the header.** Every element in the file, with its number, what the standard calls it and what it holds, searchable. Above it, the list of what in this particular file identifies the patient — which is a great deal more than the name.
7. **Take what you need.** The frame on screen as a PNG, with the window you set and nothing burned into it, or the whole header as plain text. Both are built in the page from what is already there.

## The longer version

[How to open a DCM file, and what is in one](https://abox.tools/guides/open-a-dicom-file/): What a .dcm file is and what opens one, what is on a hospital disc, why the files have no extension, how to open a DCM file in a browser, what window and level actually does, and what a scan carries about the patient besides the picture.

## Also in the box

- [Image to ICO](https://abox.tools/image-to-ico/): One picture in. Every size a browser, Windows or a Mac asks for, out.
- [Image to Data URI](https://abox.tools/image-to-data-uri/): The whole picture as one line of text. Paste it straight into CSS or HTML.
- [SVG to Image](https://abox.tools/svg-to-image/): Name the size. A vector has none of its own to lose.
- [Image to SVG](https://abox.tools/image-to-svg/): One shape, one outline. Point at whatever should not be there.

## Questions

### Is my scan uploaded anywhere?

No. The file is read, decoded and drawn by your own browser on your own hardware. There is no server side to this tool, and the page's `Content-Security-Policy` names every address it may contact — none of them belongs to this site. Unplug from the network and it still opens scans. \
\
That is worth more here than on any other page of this site. A DICOM file carries the patient's name, date of birth and hospital number in its header, so uploading one to a viewer means handing a stranger a medical record, not a picture.

### How do I open a .dcm file?

Drop it on the box at the top of this page. A `.dcm` file is a DICOM file — the format every hospital scanner writes — and this page is a DCM file opener: it reads the file in your browser and shows the scan, with nothing to install and nothing uploaded. It works the same on Windows, on a Mac, on Linux, on a Chromebook and on a phone, because the only program involved is the browser you already have. \
\
If what you were given is a disc or a folder rather than one file, drop the whole folder. A CT or an MR is one file per slice, often with no `.dcm` on the end of the names at all, and opening them together is what puts the scan back in order. The `DICOMDIR` file and the Windows viewer that came on the disc can be left where they are; neither is needed.

### What is a .dcm file, and why does double-clicking it do nothing?

It is one image, or one loop of images, from a medical scanner, in the DICOM format: a header of several hundred fields — who the patient is, which machine took it, how the pixels are laid out — followed by the pixels themselves. It is not a JPEG with a different name, which is why a photo app, an image editor and a browser tab all refuse it, and why a computer that has never had a medical viewer installed has nothing registered to open one. \
\
Not every file ending in `.dcm` is a scan, either: the same three letters are used by a disk-cataloguing program and by engine calibration files in the car industry. This page checks the file itself rather than its name, and says so plainly when what it was given is not DICOM.

### Which DICOM files can it open?

Uncompressed files in any of the three basic transfer syntaxes — implicit and explicit little endian, and the retired big-endian one — plus deflated, RLE Lossless, baseline JPEG, and JPEG Lossless, which is what the majority of CT and MR studies on a hospital disc are compressed with. \
\
It cannot decode JPEG 2000, JPEG-LS, or the MPEG and HEVC syntaxes used for video. Those need codecs that are megabytes of compiled library, and a page that downloaded one on demand would not be a page that works offline. A file in one of them still opens: the whole header is read and shown, and the picture is replaced by a line naming the codec, rather than by a broken-image icon that tells you nothing.

### What is “window and level”, and why do I need it?

A CT slice holds about four thousand distinct values and your screen shows two hundred and fifty-six greys. The window is the choice of which slice of that range gets all of them: everything below it is black, everything above it is white, and what is in between is spread across the greys. \
\
That is why the same file looks like a different scan under two settings, and why lung and bone cannot be seen at once. On a CT the numbers are Hounsfield units, which are defined absolutely — water is 0 and air is −1000 — so the named windows on this page are the same numbers a radiologist uses at a workstation. On an MR or an ultrasound there is no such scale, and the window that opens is the one the file itself asks for.

### Why does it say my measurement is in pixels?

Because that file does not say how big a pixel is. Pixel Spacing (0028,0030) is what carries that, in millimetres, and a great many ultrasound images, scanned documents and secondary captures simply do not have it. \
\
Where it is there, the measurement is in millimetres and each axis is measured with its own spacing, which matters on the images whose pixels are not square. Where it is not, the honest answer is a count of pixels, and this says so rather than picking a scale and presenting the result as a length.

### It opened my folder as several series. Why?

Because that is what is in it. A study is made of series — the scout view, then each acquisition or reconstruction — and every file says which one it belongs to in Series Instance UID (0020,000E). The dropdown is built from that rather than from the folder, which usually has all of them mixed together in one list of names. \
\
Within a series the slices are put in order by where each one sits in the patient, worked out from Image Position and Image Orientation. Instance Number is the obvious key and it is the fallback rather than the first choice: it is assigned by whatever wrote the files and does not have to run in the direction the patient does.

### What does the “what identifies the patient” list mean?

It is every field in your file that names the person the scan is of, or that narrows down who they could be, read off this file on your machine. The list comes from PS3.15 of the DICOM standard — the part that says what has to go before a dataset can be called de-identified. \
\
It is there because the thing people get wrong is not that a scan has a name in it. It is how much else it has: the birth date, the accession number, the referring physician, the institution, the scanner's serial number, and the study UIDs, which are perfect keys back into the archive that made the file. A scan that has had the name blanked out and nothing else is not anonymous. \
\
This tool only shows you. It writes nothing and changes nothing, so it cannot take any of it out.

### Can it anonymise a scan?

No, and it deliberately does not pretend to. This page reads; it has no code that writes a DICOM file. What it does is tell you exactly what is in yours, which is the part that is hard to find out and the part people are wrong about. \
\
A tool that takes the identifiers out is a separate job with a much higher bar — it has to rewrite the file without touching the pixels, replace the UIDs consistently across a whole study, and be right about the private elements some scanners hide a second copy of the name in. It is on the roadmap for this site rather than bolted onto a viewer.

### Can it open a file that has no .dcm extension, or a broken one?

Yes to both. The extension is not looked at: what is checked is the file itself. A dataset written without the usual 128-byte preamble — which is what a scan pulled straight off the network looks like — is read by working out its encoding from its first element, and the page says that is what it did. \
\
A file that ends part-way through is read as far as it goes. Everything before the damage is shown, with a note saying which byte it stopped at. That is the case a viewer is most wanted for, so throwing the whole file away over its last twelve bytes would be the wrong behaviour.

### Is this a diagnostic viewer?

No. It is not a medical device, it has not been through any regulatory assessment, and nothing here should be used to make a clinical decision. Your screen is not calibrated, the browser is not a validated rendering chain, and neither of those can be fixed from inside a web page. \
\
What it is good for is everything else people open a scan for: checking what is on a disc, pulling a slice for a teaching deck or a paper, reading a header, working out why another program refuses the file, and seeing what a scan is carrying about the person it is of.

### Does it change my file?

No. This tool only reads. There is no output file, no re-encode and no button here that writes a DICOM — what you can download is a PNG of the frame on screen and a plain-text copy of the header. Your original is untouched on your disk.

### Is it free, and do I need an account?

It is free, and there is no account, no sign-in and no trial. There is no limit on file size or on how many files you open beyond your own machine's memory. The site carries advertising, which is what pays for it; the ads are not given anything about your file.

### Does it work offline?

Yes. Load the page once, then disconnect from the internet and it keeps working. That is also the simplest way to prove nothing is being uploaded: a tool that sent your scan away to be rendered would stop the moment you unplugged.

## How the privacy claim is verifiable

- **Processing happens in this browser.** This tool processes input and results inside your browser. Its processing code does not upload them for processing. Content-Security-Policy restricts permitted resources and connection destinations, but allowed hosts can still receive requests: the policy does not inspect payloads and alone cannot prove that nothing is sent. Compare the source with the requests for the operation you use, or complete that local operation while disconnected.
- **This matters more here than on the other pages.** A DICOM file is not a picture with some metadata on it. It is a medical record with a picture in it: the patient's name, date of birth, hospital number, accession number, the referring doctor, the institution and the scanner's serial number are all fields in the header, and they travel with the file wherever it goes. Uploading one to a website to look at it means handing all of that to whoever runs the website. That is the thing this page exists not to do.
- **The reader is fourteen files in this repository.** Nothing here uses a library fetched from anywhere. `src/dicom.js` walks the file, `src/dictionary.js` knows what the tags are called, `src/pixels.js` turns the bytes back into measurements, `src/rle.js` and `src/jpeg-lossless.js` expand the two compressed forms this page can decode, and `src/window.js` maps what was measured onto the greys of your screen.
- **The identifiers are listed for you, and for nobody else.** The page prints out every field in your file that names or narrows down the person it is of, because that is a question somebody about to share a slice needs answered and no viewer answers it. It is put on the screen in front of you and goes nowhere else: there is no analytics event in this repository that carries any of it. Optional feedback reports only the tool slug, a fixed vote and a fixed reason, never the metadata read from your file.
- **It reads. It does not write.** There is no button here that changes your file, and no code that could. What you can take away is a PNG of the frame on screen and a text copy of the header, both built in the page from what is already on it. Your original is untouched on your disk, which is also the honest answer to what happens if you close the tab.
- **What Google loads, and what it is not given.** The ad and measurement scripts come from Google. Neither is handed anything about your file: not the pixels, not a thumbnail, not a name, a tag, a patient or a filename. Every line that parses, decodes or draws a scan is served from this origin and listed in the repository.
- **What the donate button loads, and what it is not given.** The "Buy me a coffee" button in the header is drawn by a script from cdnjs.buymeacoffee.com and letters itself from Google Fonts. It is a link and nothing more: it reports no visit, and it is handed nothing about you or your files. Nothing happens unless you click it, and what you would be clicking through to is somebody else's site.
- **It works offline.** Disconnect from the network and every part of this page still works. That is the simplest proof of all: a tool that sent your scan away to be rendered would stop the moment you unplugged.

**Check it yourself.** None of the above has to be taken on trust. This page is generated from the templates and config in the repository by a build script you can read and run yourself, and the result is committed to the `dist` branch — so you can diff what is served against what a build of the sources produces: https://github.com/A-Box-of-Tools/website

The files worth reading first are `config/site.toml` for the Content-Security-Policy, `src/dicom.js` for the parser that walks the file, `src/pixels.js` for the pixel decoding, `src/jpeg-lossless.js` for the codec most hospital exports use, and `src/window.js` for the window and level — none of which has a line that could reach the network.

How to open a DCM (DICOM) file

Choose the files. One .dcm file, or the whole folder off the disc — a CT or an MR is one file per slice, and dropping all of them at once is what puts the series back together. Files are read straight off your disk by the browser; nothing is sent anywhere while you do it.
Pick the series. A study usually holds several: the scout, then each acquisition. Each one is stacked in the order the scanner took it in, worked out from where each slice sits in the patient rather than from the numbering, which is not always in the same direction.
Set the window. This is the control that makes a scan readable, and the one an image editor does not have. Drag across the picture to widen the window and up or down to move its centre, or pick one of the named windows — lung, bone, brain, soft tissue — on a CT, where the units are the same on every scanner in the world.
Scroll the stack. The slider under the picture moves through the slices, and the arrow keys do the same thing once you have clicked on the picture. A multi-frame file — an ultrasound loop, an angiogram — plays with the button beside it.
Measure something. Switch to Measure and drag a line. Where the file says how far apart its pixels are, the answer is in millimetres and takes account of pixels that are not square; where the file does not say, the answer is in pixels and says so rather than inventing a scale.
Read the header. Every element in the file, with its number, what the standard calls it and what it holds, searchable. Above it, the list of what in this particular file identifies the patient — which is a great deal more than the name.
Take what you need. The frame on screen as a PNG, with the window you set and nothing burned into it, or the whole header as plain text. Both are built in the page from what is already there.

The longer version

How to open a DCM file, and what is in one What a .dcm file is and what opens one, what is on a hospital disc, why the files have no extension, how to open a DCM file in a browser, what window and level actually does, and what a scan carries about the patient besides the picture.

Questions

Is my scan uploaded anywhere?

No. The file is read, decoded and drawn by your own browser on your own hardware. There is no server side to this tool, and the page's Content-Security-Policy names every address it may contact — none of them belongs to this site. Unplug from the network and it still opens scans.

That is worth more here than on any other page of this site. A DICOM file carries the patient's name, date of birth and hospital number in its header, so uploading one to a viewer means handing a stranger a medical record, not a picture.

How do I open a .dcm file?

Drop it on the box at the top of this page. A .dcm file is a DICOM file — the format every hospital scanner writes — and this page is a DCM file opener: it reads the file in your browser and shows the scan, with nothing to install and nothing uploaded. It works the same on Windows, on a Mac, on Linux, on a Chromebook and on a phone, because the only program involved is the browser you already have.

If what you were given is a disc or a folder rather than one file, drop the whole folder. A CT or an MR is one file per slice, often with no .dcm on the end of the names at all, and opening them together is what puts the scan back in order. The DICOMDIR file and the Windows viewer that came on the disc can be left where they are; neither is needed.

What is a .dcm file, and why does double-clicking it do nothing?

It is one image, or one loop of images, from a medical scanner, in the DICOM format: a header of several hundred fields — who the patient is, which machine took it, how the pixels are laid out — followed by the pixels themselves. It is not a JPEG with a different name, which is why a photo app, an image editor and a browser tab all refuse it, and why a computer that has never had a medical viewer installed has nothing registered to open one.

Not every file ending in .dcm is a scan, either: the same three letters are used by a disk-cataloguing program and by engine calibration files in the car industry. This page checks the file itself rather than its name, and says so plainly when what it was given is not DICOM.

Which DICOM files can it open?

Uncompressed files in any of the three basic transfer syntaxes — implicit and explicit little endian, and the retired big-endian one — plus deflated, RLE Lossless, baseline JPEG, and JPEG Lossless, which is what the majority of CT and MR studies on a hospital disc are compressed with.

It cannot decode JPEG 2000, JPEG-LS, or the MPEG and HEVC syntaxes used for video. Those need codecs that are megabytes of compiled library, and a page that downloaded one on demand would not be a page that works offline. A file in one of them still opens: the whole header is read and shown, and the picture is replaced by a line naming the codec, rather than by a broken-image icon that tells you nothing.

What is “window and level”, and why do I need it?

A CT slice holds about four thousand distinct values and your screen shows two hundred and fifty-six greys. The window is the choice of which slice of that range gets all of them: everything below it is black, everything above it is white, and what is in between is spread across the greys.

That is why the same file looks like a different scan under two settings, and why lung and bone cannot be seen at once. On a CT the numbers are Hounsfield units, which are defined absolutely — water is 0 and air is −1000 — so the named windows on this page are the same numbers a radiologist uses at a workstation. On an MR or an ultrasound there is no such scale, and the window that opens is the one the file itself asks for.

Why does it say my measurement is in pixels?

Because that file does not say how big a pixel is. Pixel Spacing (0028,0030) is what carries that, in millimetres, and a great many ultrasound images, scanned documents and secondary captures simply do not have it.

Where it is there, the measurement is in millimetres and each axis is measured with its own spacing, which matters on the images whose pixels are not square. Where it is not, the honest answer is a count of pixels, and this says so rather than picking a scale and presenting the result as a length.

It opened my folder as several series. Why?

Because that is what is in it. A study is made of series — the scout view, then each acquisition or reconstruction — and every file says which one it belongs to in Series Instance UID (0020,000E). The dropdown is built from that rather than from the folder, which usually has all of them mixed together in one list of names.

Within a series the slices are put in order by where each one sits in the patient, worked out from Image Position and Image Orientation. Instance Number is the obvious key and it is the fallback rather than the first choice: it is assigned by whatever wrote the files and does not have to run in the direction the patient does.

What does the “what identifies the patient” list mean?

It is every field in your file that names the person the scan is of, or that narrows down who they could be, read off this file on your machine. The list comes from PS3.15 of the DICOM standard — the part that says what has to go before a dataset can be called de-identified.

It is there because the thing people get wrong is not that a scan has a name in it. It is how much else it has: the birth date, the accession number, the referring physician, the institution, the scanner's serial number, and the study UIDs, which are perfect keys back into the archive that made the file. A scan that has had the name blanked out and nothing else is not anonymous.

This tool only shows you. It writes nothing and changes nothing, so it cannot take any of it out.

Can it anonymise a scan?

No, and it deliberately does not pretend to. This page reads; it has no code that writes a DICOM file. What it does is tell you exactly what is in yours, which is the part that is hard to find out and the part people are wrong about.

A tool that takes the identifiers out is a separate job with a much higher bar — it has to rewrite the file without touching the pixels, replace the UIDs consistently across a whole study, and be right about the private elements some scanners hide a second copy of the name in. It is on the roadmap for this site rather than bolted onto a viewer.

Can it open a file that has no .dcm extension, or a broken one?

Yes to both. The extension is not looked at: what is checked is the file itself. A dataset written without the usual 128-byte preamble — which is what a scan pulled straight off the network looks like — is read by working out its encoding from its first element, and the page says that is what it did.

A file that ends part-way through is read as far as it goes. Everything before the damage is shown, with a note saying which byte it stopped at. That is the case a viewer is most wanted for, so throwing the whole file away over its last twelve bytes would be the wrong behaviour.

Is this a diagnostic viewer?

No. It is not a medical device, it has not been through any regulatory assessment, and nothing here should be used to make a clinical decision. Your screen is not calibrated, the browser is not a validated rendering chain, and neither of those can be fixed from inside a web page.

What it is good for is everything else people open a scan for: checking what is on a disc, pulling a slice for a teaching deck or a paper, reading a header, working out why another program refuses the file, and seeing what a scan is carrying about the person it is of.

Does it change my file?

Is it free, and do I need an account?

Does it work offline?

DICOM Viewer — a DCM file opener that runs in your browser

How the privacy claim is verifiable

Open a DCM file

What this file is

What in this file identifies the patient

How to open a DCM (DICOM) file

The longer version

Questions

How the privacy claim is verifiable

Open a DCM file

—

What this file is

What in this file identifies the patient

Every tag in the file

How to open a DCM (DICOM) file

The longer version

Also in the box

Questions