Why DICOM Is Not Just an Image
A DICOM file is far more than a picture. It is a structured object that bundles pixel data with hundreds of metadata attributes — patient identifiers, acquisition parameters, equipment details, and the display settings a radiologist used. The pixel data itself often has a much higher bit depth than an ordinary image: a CT scan stores thousands of distinct intensity values (Hounsfield units), where a JPEG can represent only 256 levels of gray per channel. Converting DICOM to JPEG or PNG therefore involves real decisions, not a simple format swap.
The most important of those decisions is windowing, and understanding it is the difference between an export that looks right and one that looks washed out or impenetrably dark.
Window and Level: Mapping Wide Data to a Narrow Screen
Because a monitor and a JPEG can only show a limited range of brightness, the DICOM viewer must map a slice of the image's full intensity range onto that visible range. The window width sets how many intensity values are spread across the available gray levels, controlling contrast. The window center (or level) sets which intensity sits in the middle, controlling brightness. A narrow window with a low center reveals soft tissue; a wide window centered higher reveals bone. Most DICOM files carry suggested window values, which this converter loads by default, but you can adjust them before exporting so the exported JPEG or PNG emphasizes exactly the anatomy you care about.
JPEG vs PNG for Medical Images
JPEG uses lossy compression: it discards information the eye is unlikely to notice, producing small files ideal for presentations, emails, and reports. The quality slider trades file size against fidelity. PNG is lossless: it preserves every pixel exactly, which matters when an image will be measured, annotated, or used for documentation where compression artifacts could be misleading. As a rule, choose PNG when accuracy is paramount and JPEG when file size and convenience win. Neither format preserves the original DICOM bit depth, so for any diagnostic re-use you should keep the source .dcm file.
Multi-Frame Objects and Cine Loops
Some DICOM objects contain many frames in a single file — ultrasound cine loops, multi-phase CT, and cardiac MR are common examples. Converting these means deciding whether you want one representative frame or the entire sequence. This tool lets you scrub to a specific frame for a single export, or render every frame and bundle them into a ZIP, naming each file with its frame number so the order is preserved.
The PHI Problem Conversion Does Not Solve
Stripping the DICOM wrapper removes the structured metadata tags, but it does not remove patient information that was burned directly into the pixels — a name overlaid on an ultrasound image, an accession number in the corner of a scanned film, or an annotation added at the scanner. Those identifiers survive into the JPEG or PNG. Before sharing converted images outside a trusted environment, run the source files through a de-identification step and visually confirm that no burned-in text remains. Converting is not the same as anonymizing.
A Practical Workflow
For most users the sequence is: load the DICOM files, set the window and level so the region of interest is clear, choose JPEG for sharing or PNG for documentation, pick the frame if the object is multi-frame, and export. If the images will leave your institution, de-identify first and review each exported image for burned-in text. Because everything runs locally, you can convert sensitive studies without the compliance risk of uploading them to a third-party website.
Bit Depth and Why the Original Still Matters
One detail worth repeating: JPEG and PNG store at most 8 bits per channel, while CT and MR data are commonly 12 or 16 bits. The conversion necessarily collapses that range through your chosen window, discarding intensity information outside it. For viewing, sharing, and teaching this is exactly what you want. For anything quantitative — re-measuring Hounsfield units, re-windowing later, or feeding an algorithm — the converted image is insufficient and you must return to the source .dcm. Treat the JPEG or PNG as a faithful snapshot of one presentation, not as the study itself, and keep the original archived.