top of page
Writer's pictureClément MORICEAU

How to Evaluate Image Quality

To evaluate the quality of a scanned image involves examining several technical and visual aspects. Here are some criteria to consider:


✨ Resolution (DPI - Dots Per Inch)


The resolution of a scanned image is measured in dots per inch (DPI). The higher the resolution, the more detailed the image. For textual documents, a resolution of 300 DPI is usually sufficient, while for images or photographs, 600 DPI or more may be required.


✨ Sharpness


Sharpness refers to the clarity of the details in the image. A well-scanned image should have sharp, distinct edges without excessive blurring. Blur can be caused by movement during scanning or incorrect settings of the scanning device.


✨ Color and Contrast


The colors of the scanned image must be true to the original. The contrast should also be sufficient to distinguish between light and dark details without losing information. Dull colors or low contrast can indicate poor scanner calibration or inappropriate settings.


✨ Presence of artifacts


Artifacts are unwanted defects in an image, such as lines, smudges, digital noise, or moiré effects. These can be the result of poor equipment quality, dust, or scratches on the scanner glass.


✨ Aspect Ratio and Distortion


The scanned image must maintain the correct proportions of the original. Distortion can occur if the document is not placed correctly on the scanner or if the scanner is incorrectly calibrated.


✨ Grayscale


For grayscale documents, it is essential that the scanned image captures a full range of grays, from deep blacks to brilliant whites, without losing detail in shadows or highlights.


✨ File size


A high-quality scanned image with high resolution and good colors will usually have a larger file size. However, it is important to find a balance between quality and file size for storage and sharing needs.


✨ Uniformity of lighting


A well-scanned image should show uniform lighting, without areas that are too bright or too dark. Variations in lighting can indicate problems with the scanner or a wrong position of the document.


✨ Color Depth


Grayscale images: Often used for textual documents, a depth of 8 bits per pixel (256 grayscales) is standard.


Color images: For high-quality color images, a color depth of 24 bits (8 bits per channel for red, green, and blue, totaling 16.7 million colors) is typical.


Medical and scientific images: Some applications require a color depth of 16 bits or more per channel to capture a wider range of details and shades.


✨ File formats


JPEG: Used for photographs and images on the web, it offers good lossy compression, but reduces the quality by eliminating some data.


TIFF: Used for scanned documents and photographs, this format can be lossless and is therefore preferred for maximum quality.


PNG: Used for images on the web where transparency is needed, it offers lossless compression.


DICOM: Standard for medical images, including X-rays, CT scans, and MRIs, this format retains the metadata needed for accurate diagnosis.


✨ Compression standards


JPEG 2000: An enhanced version of JPEG that enables lossy and lossless compression, often used in digital archives and for medical images.


HEIF (High Efficiency Image Format): Used to compress images while maintaining high quality, often used by modern smartphones.


✨ Visual Quality Standards


ISO 12647: A series of international standards for printing processes that specifies quality criteria, including color density and gamut.


sRGB and Adobe RGB: Standard color spaces used to ensure accurate color reproduction on different devices. sRGB is widely used for screens and on the web, while Adobe RGB is used in professional environments for a wider color gamut.


✨ Color Calibration and Consistency


ICC Profiles (International Color Consortium): ICC profiles are used to ensure a reproducible


Qualité des images ADOC Solutions

ADOC Solutions' teams support you in managing your document processes.



0 views0 comments

Comentarios


bottom of page