Optimized Compression for Telepathology Virtual Slides

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): We propose to investigate a method to improve the presentation of telepathology images (i.e., virtual slides) for accurate diagnoses by using image compression schemes based on information about the properties and capabilities of the human visual system. Although image compression techniques have been studied and evaluated for use in many telepathology applications, we believe that there has yet to be a thorough investigation in this area that combines state-of-the-art compression methods with vision modeling and observer performance studies that take into account the compression schemes, the display, and the clinician as a dynamic and integrated entity rather than individual parts that operate independently of one another. Through this approach we believe that we can tailor image compression schemes for specific telepathology applications thereby enabling: (1) faster delivery of images without compromising visual quality, utility or diagnostic accuracy; (2) consistently uniform visual quality with variable compression rates optimized across images, patients, and display technologies; and (3) automatic, objective image-quality control based on simulations of the human observer. Our overriding hypothesis is that it is possible to improve the presentation of telepathology images for accurate diagnoses by tailoring image compression schemes and displays that are based on information about the capabilities and limitations of the human visual system. Once we have confirmed the utility of this approach with telepathology images, it can easily be generalized to other telemedicine applications that use digital color images for store-and-forward transmission such as teledermatology, as well as real-time video telemedicine applications (e.g., telepsychiatry). We hypothesize that it is possible to improve the presentation of telepathology images for accurate diagnoses using image compression schemes that are based on information about the capabilities and limitations of the human visual system. We propose two specific aims to test this hypothesis. An underlying long-term goal is to demonstrate the utility of a visual discrimination model (VDM) for predicting observer performance (i.e., detection and/or discrimination of lesions in medical images) as a function of compression in a variety of both store-forward and real-time video telemedical applications. The achievement of this goal will ensure that as more compression schemes are developed or as they are applied to unique imaging applications, time and personnel intensive Receiver Operating Characteristic (ROC) experiments (the generally preferred method for evaluating diagnostic performance) can either be avoided or reduced significantly in scope while still assessing the effects of a given compression scheme or level of compression on diagnostic performance.