Tree-structured vector quantization of CT chest scans: Image quality and diagnostic accuracy

Journal Article · Wed Dec 01 04:00:00 EST 1993 · IEEE Transactions on Medical Imaging (Institute of Electrical and Electronics Engineers); (United States)

DOI:https://doi.org/10.1109/42.251124· OSTI ID:5079548

Cosman, P C; Tseng, C; Gray, R M ^[1]; Olshen, R A; Moses, L E ^[2]; Davidson, H C; Bergin, C J ^[3]; Riskin, E A ^[4]

---

1. Stanford Univ., Stanford, CA (United States). Dept. of Electrical Engineering
2. Stanford Univ. School of Medicine, Stanford, CA (United States). Dept. of Statistics
3. Stanford Univ., Stanford, CA (United States). Dept. of Diagnostic Radiology and Nuclear Medicine
4. Univ. of Washington, Seattle, WA (United States). Dept. of Electrical Engineering

The quality of lossy compressed images is often characterized by signal-to-noise ratios, informal tests of subjective quality, or receiver operating characteristic (ROC) curves that include subjective appraisals of the value of an image for a particular application. The authors believe that for medical applications, lossy compressed images should be judged by a more natural and fundamental aspect of relative image quality: their use in making accurate diagnoses. They apply a lossy compression algorithm to medical images, and quantify the quality of the images by the diagnostic performance of radiologists, as well as by traditional signal-to-noise ratios and subjective ratings. The study is unlike previous studies of the effects of lossy compression in that they consider non-binary detection tasks, simulate actual diagnostic practice instead of using paired tests or confidence rankings, use statistical methods that are more appropriate for non-binary clinical data than are the popular ROC curves, and use low-complexity predictive tree-structured vector quantization for compression rather than DCT-based transform codes combined with entropy coding. Their diagnostic tasks are the identification of nodules (tumors) in the lungs and lymphadenopathy in the mediastinum from computerized tomography (CT) chest scans. For the image modality, compression algorithm, and diagnostic tasks they consider, the original 12 bit per pixel (bpp) CT image can be compressed to between 1 bpp and 2 bpp with no significant changes in diagnostic accuracy.

• OSTI ID: 5079548
• Journal Information: IEEE Transactions on Medical Imaging (Institute of Electrical and Electronics Engineers); (United States), Journal Name: IEEE Transactions on Medical Imaging (Institute of Electrical and Electronics Engineers); (United States) Vol. 12:4; ISSN 0278-0062; ISSN ITMID4
• Country of Publication: United States
• Language: English

Related Subjects

550601* -- Medicine-- Unsealed Radionuclides in Diagnostics
62 RADIOLOGY AND NUCLEAR MEDICINE
ACCURACY
BODY
BODY AREAS
CHEST
COMPUTERIZED TOMOGRAPHY
DIAGNOSIS
DIAGNOSTIC TECHNIQUES
DISEASES
IMAGE PROCESSING
LUNGS
NEOPLASMS
ORGANS
PROCESSING
RESPIRATORY SYSTEM
TOMOGRAPHY