Michael A. McFadden

Comparison of the low-contrast detectability of a screen-film system and third generation computed radiography.

Contrast-detail (CD) analysis was used to compare the low-contrast detectability of computed radiography (CR) and screen-film (SF) as applied to the task of adult chest radiography. A phantom was constructed and imaged using the same exposure factors throughout all experiments. Within-observer variance, between-observer variance, and image sample variance were calculated and used to estimate the standard error for each experiment. The results of these CD experiments agreed with the predictions of the Rose model. Observers performed equally well for low-contrast target detection using CR and SF.

Teleradiology using switched dialup networks

A dialup wide-bandwidth digital teleradiology system is discussed. A laser film digitizer and a gray scale display system are used at Irwin and Munson Army Community Hospitals to digitize radiographic films and display digital images. A laser film printer at the University of Kansas Medical Center generates a film hardcopy of the transmitted digital data, and an interactive gray scale display is used to review the digital image data. The communication system consists of dialup switched multiple 56000 b/s digital channels, transmitting digital image data in parallel. Conventional radiographic films, multiformat camera films, and laser printed films from multimodality imaging systems have been successfully digitized, transmitted, and laser film recorded or gray scale displayed. It was found that the system provided clinically acceptable image quality reproductions. >

The effects of lossy compression on the detection of subtle pulmonary nodules.

We examined the ability of radiologists to detect pulmonary nodules in computed radiographic (CR) chest images subjected to lossy image compression. Low-contrast 1-cm diameter targets simulating noncalcified pulmonary nodules were introduced into clinical images and presented to ten radiologists in a series of two-alternative forced-choice (2AFC) observer experiments. The percentages of correct observer responses obtained while viewing noncompressed images (1:1) were compared with those obtained for the same images compressed 7:1, 16:1, 44:1, and 127:1. The images were compressed using a standard full-frame discrete cosine transform (DCT) technique. The degree of compression was determined by quantizing Fourier components in various frequency channels and then Huffman encoding the result. The data show a measurable decline in performance for each compression ratio. Through signal-to-noise ratio (SNR) analysis, we found that the reduction in performance was due primarily to the compression algorithm that increased image noise in the frequency channels of the signals to be detected.

Contrast‐detail analysis of image degradation due to lossy compression

A contrast-detail (CD) experiment was performed to study the effect of lossy compression on computed radiographic (CR) images. Digital CR images of a phantom were compressed by quantizing the full-frame discrete cosine transform and Huffman encoding the result. Since low-contrast detectability is directly linked to an important radiological task, namely, the detection of noncalcified pulmonary nodules in adult chest radiographs, the goal of the study was to quantify any loss in low-contrast detectability due to compression. Compression ratios varied significantly among compressed images, despite the use of fixed compression parameters; detectability could be specified by a single parameter of a CD curve; there was no significant reduction in detectability for an average compression ratio of 11:1; and, there was a statistically significant degradation in detectability for an average compression ratio of 125:1.

Image data compression using a new floating-point digital signal processor

A new dual-ported, floating-point, digital signal processor has been evaluated for compressing 512 and 1,024 digital radiographic images using a full-frame, two-dimensional, discrete cosine transform (2D-DCT). The floating point digital signal processor operates at 49.5 million floating point instructions per second (MFLOPS). The level of compression can be changed by varying four parameters in the lossy compression algorithm. Throughput times were measured for both 2D-DCT compression and decompression. For a 1,024×1,024×10-bit image with a compression ratio of 316∶1, the throughput was 75.73 seconds (compression plus decompression throughput). For a digital fluorography 1,024×1,024×8-bit image and a compression ratio of 26∶1, the total throughput time was 63.23 seconds. For a computed tomography image of 512×512×12 bits and a compression ratio of 10∶1 the throughput time was 19.65 seconds.

Assessment of low-contrast detectability for compressed digital chest images

The ability of human observers to detect low-contrast targets in screen-film (SF) images, computed radiographic (CR) images, and compressed CR images was measured using contrast detail (CD) analysis. The results of these studies were used to design a two- alternative forced-choice (2AFC) experiment to investigate the detectability of nodules in adult chest radiographs. CD curves for a common screen-film system were compared with CR images compressed up to 125:1. Data from clinical chest exams were used to define a CD region of clinical interest that sufficiently challenged the observer. From that data, simulated lesions were introduced into 100 normal CR chest films, and forced-choice observer performance studies were performed. CR images were compressed using a full-frame discrete cosine transform (FDCT) technique, where the 2D Fourier space was divided into four areas of different quantization depending on the cumulative power spectrum (energy) of each image. The characteristic curve of the CR images was adjusted so that optical densities matched those of the SF system. The CD curves for SF and uncompressed CR systems were statistically equivalent. The slope of the CD curve for each was - 1.0 as predicted by the Rose model. There was a significant degradation in detection found for CR images compressed to 125:1. Furthermore, contrast-detail analysis demonstrated that many pulmonary nodules encountered in clinical practice are significantly above the average observer threshold for detection. We designed a 2AFC observer study using simulated 1-cm lesions introduced into normal CR chest radiographs. Detectability was reduced for all compressed CR radiographs.

Phosphor plate mammography: contrast studies and clinical experience

Mammography and accurate microcalcification detection require very good spatial resolution. We have compared the diagnostic capabilities of reduced-exposure, third-generation, 5 cycles/mm computed radiography (CR) phosphor plates with conventional screen-film in 67 patients. No difference in diagnostic accuracy was detected. The digital characteristics of storage phosphor plates erabled us to study the relationship between contrast and spatial resolution. We developed a computer program to identify a single 100 μm pixel in a digital image and assign various gray levels to that pixel. Using this model, we determined that, for our 5 cycles/mm CR system, the imaged contrast of a 100 μm object was 62% of the original contrast. Current 5 cycles/mm phosphor plate systems cannot adequately detect microcalcifications that approximate 100 μm or smaller unless a magnification technique is used.

Contrast-detail analysis of the effect of image compression on computed tomographic images

Three compression algorithms were compared by using contrast-detail (CD) analysis. Two phantoms were designed to simulate computed tomography (CT) scans of the head. The first was based on CT scans of a plastic cylinder containing water. The second was formed by combining a CT scan of a head with a scan of the water phantom. The soft tissue of the brainwas replaced by a subimage containing only water. The compression algorithms studied were the full-frame discrete cosine (FDCT) algorithm, the Joint Photographic Experts Group (JPEG) algorithm, and a wavelet algorithm. Both the wavelet and JPEG algorithms affected regions of the image near the boundary of the skull. The FDCT algorithm propagated false edges throughout the region interior to the skull. The wavelet algorithm affected the images less than the other compression algorithms. The presence of the skull especially affected observer performance on the FDCT compressed images. All of the findings demonstrated a flattening of the CD curve for large lesions. The results of a compression study using lossy compression algorithms is dependent on the characteristics ofthe image and the nature of the diagnostic task. Because of the high density bone of the skull, head CT images present a much more difficult compression problem than chest x-rays. We found no significant differences among the CD curves for the tested compression algorithms. Key Words: Image compression, contrast-detail analysis.

Further investigation into the effects of lossy compression using free-response methodology

Using free-response methodology, we measured the comparative performance of trained observers for the task of detecting simulated pulmonary nodules in computed radiographic chest images that were compressed using the full-frame discrete cosine transform algorithm. six observers read fifty-one images containing a total of 372 simulated lesions of size ranging from 8 mm to 12 mm and with six different contrasts. The images were compressed to an average of 15:1 with the same parameters that were used in an earlier two-alternative forced- choice analysis. The results showed this level of compression did not increase the number of false=positive calls per image. also, observers tended to ignore the lowest contrast nodules in all images. At low contrast we expect compression to have the greatest effect. Therefore, overall performance was not degraded by the compression process, although performance was compromised at very low contrast.