学术文献库

The need for fast diagnostic image viewing in zero footprint web applications and the ever increasing image sizes for new modalities such as digital pathology have painfully brought to light that the currently available image compression methods fall short. JPEG2000 delivers the image quality required for medical grade viewing, but is supported on fewer than 10% of desktop web browsers installed today (caniuse.com) and even then it does not support the high bit depth images required by medical applications. JPEG2000's high computational complexity and inability to do fast compression and viewing of images undoubtedly contributed to its lack of adoption. The venerable JPEG standard is supported in all installed web browsers today, and allows for fast viewing and compression, but it cannot provide medical grade image quality, lossless compression, or high bit depths. To remedy the situation medical image web applications need to take full control of the image path by implementing the image decompression in the application itself, instead of relying on web browser implementations. This will give users and manufacturers the confidence that medical images will be displayed in their full intended fidelity! This paper introduces a simple, fast, yet efficient image compression method that can be implemented in zero footprint web applications to provide fast medical grade image viewing on today's web browser installed base, on the premise and in the cloud. Due to its efficiency it is very suitable for fast and affordable compression and viewing of very large images, and is used as such by Philips in the iSyntax file format in use in its digital pathology products today.