Bitmap data and raster data is the format for displaying images as a cluster of pixels.

Images are digitized when photographed by a digital camera or scanned. Sampling and quantizing are required for digitizing analog content. Sampling is the breakdown of the image into pixel grids while quantizing is the addition of discrete values to the contrasting densities and color tones of the pixels. The resolution is determined in this process by adjusting the degree of sampling, and monochrome to full color images are reproduced by adjusting the amount of information contained in each pixel (color depth). The theoretical data size of a digitized image is calculated by multiplying the vertical pixel count, the horizontal pixel count and the color depth, and then dividing the product by 8 (the unit will be in bytes). Therefore, the data size will increase if the resolution or color depth increases, and decrease if these decrease.

Raster data is used in scanning-type input and output devices to clarify the way in which a row of pixels in a bitmap data is treated as single unit.

The resolution of an image data refers to how many pixels are sampled in 1 inch (25.4mm) and is expressed as dots per inch (dpi). Generally, images with a resolution of around 350dpi are used for printing while this is kept down to around 72dpi for Web use.

A variety of file formats can be used to record image data on a medium, but an appropriate format must be selected according to its use and purpose, from uncompressed formats such as TIFF and BMP (or those with a lossless compression option) to JPEG and other lossy compression formats. Whichever format is used, the file can be used in paint-style image processing tools such as Photoshop from Adobe Systems or the free software, GIMP. However, care should be taken in how the images are used because there is no way of recovering the image quality once it has been degraded through editing and retouching.

The file size can vary depending on the size of the header section and the color palette or the amount of padding data, and may therefore be different from the theoretical value even when uncompressed.