Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “object detection”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 325 records · Page 18Linked to original sources

Object-related activity revealed by functional magnetic resonance imaging in human occipital cortex.

The stages of integration leading from local feature analysis to object recognition were explored in human visual cortex by using the technique of functional magnetic resonance imaging. Here we report evidence for object-related activation. Such activation was located at the lateral-posterior aspect of the occipital lobe, just abutting the posterior aspect of the motion-sensitive area MT/V5, in a region termed the lateral occipital complex (LO). LO showed preferential activation to images of objects, compared to a wide range of texture patterns. This activation was not caused by a global difference in the Fourier spatial frequency content of objects versus texture images, since object images produced enhanced LO activation compared to textures matched in power spectra but randomized in phase. The preferential activation to objects also could not be explained by different patterns of eye movements: similar levels of activation were observed when subjects fixated on the objects and when they scanned the objects with their eyes. Additional manipulations such as spatial frequency filtering and a 4-fold change in visual size did not affect LO activation. These results suggest that the enhanced responses to objects were not a manifestation of low-level visual processing. A striking demonstration that activity in LO is uniquely correlated to object detectability was produced by the "Lincoln" illusion, in which blurring of objects digitized into large blocks paradoxically increases their recognizability. Such blurring led to significant enhancement of LO activation. Despite the preferential activation to objects, LO did not seem to be involved in the final, "semantic," stages of the recognition process. Thus, objects varying widely in their recognizability (e.g., famous faces, common objects, and unfamiliar three-dimensional abstract sculptures) activated it to a similar degree. These results are thus evidence for an intermediate link in the chain of processing stages leading to object recognition in human visual cortex.

Adult↗

Positive prostate-specific antigen circulating cells detected by reverse transcriptase-polymerase chain reaction does not imply the presence of prostatic micrometastases.

OBJECTIVES: Detection of circulating tumor cells may improve the preoperative local staging of prostate cancers. The aim of this study was to perform enhanced reverse transcriptase-polymerase chain reaction (RT-PCR) of prostate-specific antigen (PSA) mRNA to define the predictive value of PSA-positive circulating cells in a large series of patients. METHODS: The study included 46 patients with Stage T1 to T2 prostate cancer, 94 with benign prostatic hyperplasia (BPH), and 51 (including 9 women) with nonprostatic disease. PSA-positive cells from peripheral blood samples were detected by Southern blot analysis of the RT-PCR products. Original oligonucleotide primers were defined to exclusively detect the three PSA mRNA splices. RESULTS: Circulating PSA-positive cells were observed in 8 (8.5%) of 94 patients with BPH, 10 (22%) of 46 with Stage T1 to T2 prostate cancer, and 9 (17.6%) of 51 with nonprostatic disease. The detection rate of PSA-positive circulating cells was significantly increased in patients with prostate cancer versus patients with BPH (P = 0.03). Among clinically localized prostate cancers with a Gleason score less than 8, a correlation was observed between PSA-positive circulating cells and Stage pT3 cancer (P = 0.038), capsular penetration (P = 0.04), and a positive margin (P = 0.038). The specificity of the assay for Stage pT3 cancer detection was 84.6%, with a positive predictive value of 60%. CONCLUSIONS: Although RT-PCR assay may have a role in preoperative local staging, this study demonstrated the absence of tissue and tumor specificity of PSA-positive circulating cells, accounting for the weak positive predictive value of this technique.

Aged↗

A real-time polymerase chain reaction assay for detection of Pneumocystis from bronchoalveolar lavage fluid.

Pneumocystis jiroveci is an important cause of pneumonia in immunocompromised individuals. This organism cannot be cultured, and therefore, diagnosis relies on microscopic identification of the organism using stains or antibodies. Although simple, these tests are insensitive and require expertise for accurate interpretation. We developed a real-time polymerase chain reaction (PCR) assay that provides sensitive and objective detection of Pneumocystis from bronchoalveolar lavage fluid. Primers and fluorescence resonance energy transfer probes were developed that target the cdc2 gene of P. jiroveci. Assay sensitivity is 6 copies of target per microliter of sample. No cross-reactivity occurs with other pathogens, and the PCR assay has a 21% increase in clinical sensitivity as compared with Calcofluor white staining. The real-time PCR assay provides a sensitive, rapid, and objective method for the detection of Pneumocystis from bronchoalveolar lavage fluid.

Benzenesulfonates↗

Automated fetal head detection and measurement in ultrasound images by iterative randomized Hough transform.

An image-processing and object-detection method was developed to automate the measurements of biparietal diameter (BPD) and head circumference (HC) in ultrasound fetal images. The heads in 214 of 217 images were detected by an iterative randomized Hough transform. A head was assumed to have an elliptical shape with parameters progressively estimated by the iterative randomized Hough transform. No user input or size range of the head was required. The detection and measurement took 1.6 s on a personal computer. The interrun variations of the algorithm were small at 0.84% for BPD and 2.08% for HC. The differences between the automatic measurements and sonographers' manual measurements were 0.12% for BPD and -0.52% for HC. The 95% limits of agreement were -3.34%, 3.58% for BPD and -5.50%, 4.45% for HC. The results demonstrated that the automatic measurements were consistent and accurate. This method provides a valuable tool for fetal examinations.

Algorithms↗

Correlation filter: an accurate approach to detect and locate low contrast character strings in complex table environment.

Correlation has been used extensively in object detection field. In this paper, two kinds of correlation filters, Minimum Average Correlation Energy (MACE) and Extended Maximum Average Correlation Height (EMACH), are applied as adaptive shift locators to detect and locate smudgy character strings in complex tabular color flight coupon images. These strings in irregular tabular coupon are computer-printed characters but of low contrast and could be shifted out of the table so that we cannot detect and locate them using traditional algorithms. In our experiment, strings are extracted in the preprocessing phase by removing background and then based on geometric information, two correlation filters are applied to locate expected fields. We compare results from two correlation filters and demonstrate that this algorithm is a high accurate approach.

Journal Article↗

Use of multiple data streams to conduct Bayesian biologic surveillance.

INTRODUCTION: Emergency department (ED) records and over-the-counter (OTC) sales data are two of the most commonly used sources of data for syndromic surveillance. The majority of detection algorithms monitor these data sources separately and either do not combine them or combine them in an ad hoc fashion. This report outlines a new causal model that combines the two data sources coherently to perform outbreak detection. OBJECTIVES: This report describes the extension of the Population-wide Anomaly Detection and Assessment (PANDA) Bayesian biologic surveillance algorithm to combine information from multiple data streams. It also outlines the assumptions and techniques used to make this approach scalable for real-time surveillance of a large population. METHODS: A causal Bayesian network model used previously was extended to incorporate evidence from daily OTC sales data. At the level of individual persons, the actions that result in the purchase of OTC products and in admission to an ED were modeled. RESULTS: Preliminary results indicate that this model has a tractable running time consisting of 209 seconds for initialization and approximately 4 seconds for every hour's worth of ED data, as measured on a Pentium-4 three-Gigahertz machine with two Gigabytes of RAM. CONCLUSION: Preliminary results for surveillance using a new Bayesian algorithm that models the interaction between ED and OTC data are positive regarding the run time of the algorithm.

Algorithms↗

Detection of intravascular epidural catheters using 2-chloroprocaine. Influence of local anesthetic dose and nalbuphine premedication.

BACKGROUND AND OBJECTIVES: Detecting the intravascular placement of epidural catheters is essential to avoid local anesthetic toxicity. Small doses of intravenous local anesthetics produce changes in sensorium and are often used to test newly placed epidural catheters. Many parturients receive nalbuphine for analgesia prior to epidural catheter placement. This study examines how nalbuphine premedication influences symptoms following intravenous 2-chloroprocaine. METHODS: Thirty-one volunteers were randomized to receive premedication with placebo or 0.15 mg/kg of nalbuphine intravenously. Starting 10 minutes later, intravenous injections of 0.0, 0.3, 0.6, 0.9, 1.2, and 1.5 mg/kg of 2-chloroprocaine were given in random order at 10 minute intervals. After each injection, volunteers were asked to report changes in hearing, taste, or other symptoms. RESULTS: Symptoms reported (in decreasing order of frequency) were auditory changes, taste changes, dizziness/ lightheadedness, tingling in the extremities, and visual changes. The probability that volunteers will report symptoms is proportional to the dose of 2-chloroprocaine administered (auditory symptoms, P < .001; taste symptoms, P = .01; any symptoms, P < .001). Nalbuphine-premedicated volunteers were more likely to report symptoms (auditory symptoms, P = .004; taste symptoms, P = .004; any symptoms, P = .02). A dose of at least 1.5 mg/kg appears to be necessary to produce a 90% probability that patients will report symptoms when they receive 2-chloroprocaine intravenously. CONCLUSIONS: This study suggests that patients who receive nalbuphine for analgesia prior to epidural placement will be more likely to report symptoms after receiving intravenous 2-chloroprocaine during epidural test dosing.

Adult↗

Radiopacity of impression materials.

The radiographic densitites of twenty-six impression materials were measured and the values expressed as an equivalent thickness of aluminum. Under simulated clinical conditions, only ten of the materials tested could be consistently distinguished from the bone structure in periapical radiographs. It was far more dificult to detect objects with beveled margins than objects of uniform thickness. The minimum radiopacity of a material which could be detected in the periodontal tissue on a periapical radiograph was estimated.

Absorptiometry, Photon↗

Comparison of full-field digital mammography to screen-film mammography with respect to contrast and spatial resolution in tissue equivalent breast phantoms.

To determine if the improved contrast resolution of full-field digital mammography (FFDM) with reduced spatial resolution allows for superior or equal phantom object detection compared with screen-film mammography (SFM). Tissue equivalent breast phantoms simulating an adipose to glandular ratio of 50/50,30/70, and 20/80 were imaged according to each manufacturers' recommendation with four full-field digital mammography units (Fuji, Sectra, Fischer, and General Electric) and a screen-film mammography unit (MammoMatII 2000, Siemens, Munich, Germany). A total of 20 images were obtained in both hard- and soft-copy formats. For the purpose of soft-copy display, the screen-film hard-copy images were digitized with a 50 microm micron scanner. Six radiologists, experts in breast imaging, and three physicists, experts in scoring mammography phantoms, participated in a reader study where each reader scored each phantom for visibility of line-pairs and for 24 objects (fibers, clusters of specks, and masses). The data were recorded, entered into a database, and analyzed by a mixed-effect model. The limiting spatial resolution in line-pairs per millimeter visible with the digital units was less, regardless of display modality used, than that provided by the screen-film unit. The difference was statistically significant for the General Electric (p < 0.01) and Fuji digital mammography units (p = 0.03). With respect to the number of visible objects, a statistically significant higher number could be detected with the screen-film unit as compared to the Fischer (p < 0.01) and Sectra (p < 0.01) digital mammography units, but there was no significant difference between the other digital units and screen film. Overall, there was significantly better performance on the 50/50 phantom than with the 30/70 and 20/80 phantoms (p = 0.01, p < 0.01) for object visibility. For the digital mammography units, soft-copy display performed better than hard-copy display for the Fischer and Sectra images, but worse for Fuji and General Electric. In addition, soft-copy display of digitized screen-film images was significantly better than hard-copy display (p =0.02) of the original screen films for object visibility, but worse for spatial resolution. The higher contrast resolution of the FFDM units tested did not result in improved detection of line-pair resolution or objects in the phantoms tested versus screen-film mammography. The phantom performance of a digital mammography unit seems to be influenced by the type of detection task (line-pair resolution versus object visibility), the display modality (soft-copy versus hard-copy) chosen to score the phantoms, and the parenchymal pattern composition of the phantom.

Breast Neoplasms↗

[Detection and purification of cytokine secreting T lymphocytes in allogeneic reaction and a preliminary study on its clinical significance].

OBJECTIVE: Detection of cytokine secreting T lymphocytes after allogeneic peripheral blood mononuclear cell (PBMNC) stimulation was carried out and its clinical significance discussed so as to explore a new approach to study allogeneic reactive T lymphocytes. METHODS: Cytokine secretion assay (CKSA) was applied to detect T lymphocytes secreting interferon gamma (IFNgamma), interleukin-4 (IL-4) and IL-10 at single cell level in human mixed lymphocytes reaction. T cells secreting IFNgamma from PBMNC were detected in patients with acute graft versus host disease (aGVHD). RESULTS: In comparison with T cells secreting IL-4 and IL-10 which were (0.12 +/- 0.03)% and (0.10 +/- 0.03)%, respectively, a sizable proportion of T cells secreting IFNgamma could be detected (1.12 +/- 0.13)%. Preliminary result indicated that frequency of T cells secreting IFNgamma correlated with the occurrence of aGVHD. CONCLUSIONS: It is feasible to detect T lymphocytes secreting IFNgamma after allogeneic PBMNC and to apply CKSA technique for clinical research.

Cells, Cultured↗

High rate of detection of primary aldosteronism, including surgically treatable forms, after 'non-selective' screening of hypertensive patients.

BACKGROUND: Wide testing of the aldosterone : renin ratio among hypertensive individuals has revealed primary aldosteronism to be common, with most patients normokalaemic. Some investigators, however, have reported aldosterone-producing adenoma to be rare among patients so detected. OBJECTIVE: To test the hypothesis that differences among reported studies in the rate of detection of aldosterone-producing adenoma (as opposed to bilateral adrenal hyperplasia) reflect differences in the procedures used for diagnosis of primary aldosteronism, and the methods used to identify aldosterone-producing adenomas. METHODS: In the newly established Princess Alexandra Hospital Hypertension Unit (PAHHU), we used procedures developed by Greenslopes Hospital Hypertension Unit (which reports that more than 30% of patients with primary aldosteronism have aldosterone-producing adenomas) to diagnose primary aldosteronism and determine the subtype. All patients with an increased aldosterone : renin ratio (measured after correction for hypokalaemia and while the patient was not receiving interfering medications) underwent fludrocortisone suppression testing to confirm or exclude primary aldosteronism; if they were positive, they underwent genetic testing to exclude glucocorticoid-remediable aldosteronism before adrenal venous sampling was used to differentiate lateralizing from bilateral primary aldosteronism. RESULTS: This approach allowed PAHHU to diagnose, within 2 years, 54 patients [only seven (13%) hypokalaemic] with primary aldosteronism. All tested negative for glucocorticoid-remediable aldosteronism. Aldosterone production was lateralized to one adrenal in 15 patients (31%; only six hypokalaemic) and was bilateral in 34 (69%; all normokalaemic) of 49 patients who underwent adrenal venous sampling. Among patients with lateralizing adrenal hyperplasia, computed tomography revealed an ipsilateral mass in only six and a contralateral lesion in one. Fourteen patients underwent unilateral adrenalectomy, which cured the hypertension in seven and improved it in the remainder. In patients with bilateral primary aldosteronism, hypertension responded to spironolactone (12.5-50 mg/day) or amiloride (2.5-10 mg/day). CONCLUSION: When performed with careful regard to confounding factors, measurement of the aldosterone : renin ratio in all hypertensive individuals, followed by fludrocortisone suppression testing to confirm or exclude primary aldosteronism and adrenal venous sampling to determine the subtype, can result in the detection of significant numbers of patients with specifically treatable or potentially curable hypertension.

Adenoma↗

Detection of faecal leucocytes & erythrocytes from stools of cholera patients suggesting an evidence of an inflammatory response in cholera.

BACKGROUND & OBJECTIVES: Detection of faecal leucocytes and RBCs in stool samples of cholera patients has been reported in a small number of studies. This study extends these observations by examining stool samples of cholera patients in Calcutta. METHODS: Out of 1562 diarrhoeal stool samples, Vibrio cholerae was isolated in 266 cases. Stool samples obtained were examined microscopically within two hours of collection. RBCs and faecal leucocytes were examined by normal saline and methylene blue stain. Stool culture was performed using selective and differential media for isolation of V. cholerae. RESULTS: Among 266 cholera patients, RBCs was detected in 58 per cent and faecal leucocytes in 88 per cent respectively. The extent of the changes correlated with clinical severity. INTERPRETATION & CONCLUSION: This study showed the presence of RBCs and faecal leucocytes in stools of patients of cholera caused by V. cholerae 01 and 0139 which indicates some inflammatory changes in the gut mucosa. Further study is required to elucidate the inflammatory mechanism involved in the underlying process(es).

Cell Separation↗

Building models of animals from video.

This paper argues that tracking, object detection, and model building are all similar activities. We describe a fully automatic system that builds 2D articulated models known as pictorial structures from videos of animals. The learned model can be used to detect the animal in the original video--in this sense, the system can be viewed as a generalized tracker (one that is capable of modeling objects while tracking them). The learned model can be matched to a visual library; here, the system can be viewed as a video recognition algorithm. The learned model can also be used to detect the animal in novel images--in this case, the system can be seen as a method for learning models for object recognition. We find that we can significantly improve the pictorial structures by augmenting them with a discriminative texture model learned from a texture library. We develop a novel texture descriptor that outperforms the state-of-the-art for animal textures. We demonstrate the entire system on real video sequences of three different animals. We show that we can automatically track and identify the given animal. We use the learned models to recognize animals from two data sets; images taken by professional photographers from the Corel collection, and assorted images from the Web returned by Google. We demonstrate quite good performance on both data sets. Comparing our results with simple baselines, we show that, for the Google set, we can detect, localize, and recover part articulations from a collection demonstrably hard for object recognition.

Algorithms↗

[Detection of occult tumor cells in resected lymph nodes of patients with stage I carcinoma and its clinico-pathological significance].

OBJECTIVE: Detection of tumor micrometastases in resected lymph nodes of cancer patients to obtain more reliable information concerning nodal status and its clinicopathological importance. METHODS: Paraffin blocks of 3,715 resected lymph nodes from 350 Stage I cancer patients, suffering from NSCLC (n = 94), breast cancer (n = 112), esophageal cancer (n = 115), and vulvar cancer (n = 29), were re-examined by immunohisto-chemical staining, using monoclonal anti-cyokeratins (AE1/AE3), anti-EMA, and polyclonal anti-keratins antibodies. RESULTS: Occult nodal metastases was observed in 113 of 350 cancer patients (32.5%), and in 203 of 3,715 lymph nodes examined. The positive rates both in patients and in lymph nodes were higher in NSCLC than in others. Occult nodal metastasis was seen in 58% of pulmonary squamous-cell carcinomas and 53.8% of adenocarcinomas, while it was seen in 22.5% of esophageal and 10.3% of vulval squamous-cell carcinomas, and in 27.7% of breast adenocarcinomas (P < 0.05). Follow up of a fraction of breast cancer patients showed that the prognosis of patients with positive nodes was worse than that in patients with negative nodes (P < 0.05). CONCLUSION: The data suggested that the immunohistochemical technique can significantly facilitate the detection of micrometastatic tumor cells in lymph nodes. The frequency of occult lymph node metastasis may have significant impact on the prognosis of cancer patients.

Adenocarcinoma↗

Detection of disseminated tumor cells in mediastinoscopic lymph node biopsies and lymphadenectomy specimens of patients with NSCLC by quantitative RT-PCR.

OBJECTIVE: Detection of disseminated tumor cells in mediastinoscopic biopsies could improve staging and might be helpful concerning indications for neoadjuvant therapy regimens. This prospective study was performed to evaluate a simple and observer-independent polymerase chain reaction (PCR)-based method for the detection of disseminated tumor cells in regional lymph nodes. METHODS: Lymph nodes of 32 consecutive patients without neoadjuvant therapy were removed by systematic lymphadenectomy during resection of primary NSCLC. One hundred of these lymph nodes were cut into two equal halves which were examined using either routine histopathology or quantitative reverse transcriptase PCR (qRT-PCR). qRT-PCR amplification of cytokeratin 19 (CK19) transcripts was applied for the detection of tumor cell-specific RNA. We differentiated between illegitimate marker gene transcription and cancer-specific expression by using a cut-off value that was obtained from the analysis of 18 lymph nodes of patients with benign lung diseases. Subsequent to the evaluation of qRT-PCR, a pilot project with five additional patients was conducted to examine 19 mediastinoscopic biopsies, which were cut into two equal halves and proceeded as described above. RESULTS: Ninety-four (94%) lymph nodes were tumor-free by histopathology. qRT-PCR detected disseminated tumor cells in 26 (28%) of these lymph nodes. All of the remaining six lymph nodes that were judged by the pathologist to contain tumor cells exhibited CK19 transcripts. Twenty-three patients had a pN0 status. qRT-PCR detected disseminated tumor cells in 13 (56%) of these pN0 patients. The mediastinoscopic biopsies showed disseminated tumor cells in four (21%) out of 19 histopathologically tumor-free samples. CONCLUSIONS: CK19 qRT-PCR is a sensitive and specific tools for the detection of disseminated tumor cells in regional lymph nodes of patients with operable NSCLC. Further studies are required to asses if this molecular method might improve mediastinoscopic staging.

Aged↗

Detection of the residual lumen of intracranial aneurysms immediately after coil embolization by three-dimensional digital subtraction angiographic virtual endoscopic imaging.

OBJECTIVE: Detection of a small residual lumen after coil embolization is often difficult because of the coil mass and the overlap of the cerebral arteries. The purpose of this study was to assess the usefulness of virtual endoscopic (VE) analysis of three-dimensional digital subtraction angiographic (DSA) images for evaluation of aneurysmal occlusion immediately after the procedure. METHODS: Twenty-seven intracranial aneurysms were treated with coil embolization using a three-dimensional DSA system. Biplane and rotational DSA scanning was performed before and immediately after the procedures. VE images were obtained at a separate workstation, after transfer of the rotational images. Two-dimensional (2D) DSA images and VE images obtained after the procedure were assessed with respect to aneurysmal occlusion. Morphological outcomes and other factors, including location, size, volumetric ratio (coil volume/aneurysm volume), and residual sites, were also evaluated. RESULTS: Seven aneurysms were evaluated as complete occlusion (CO) on both 2D DSA images and VE images. Twelve aneurysms exhibited residual lumina on both 2D DSA images and VE images. Five aneurysms were evaluated as CO on 2D DSA images and as incomplete occlusion on VE images. There were no recurrences among the aneurysms that were evaluated as CO on VE images. Two of five aneurysms that were evaluated as CO on 2D DSA images and as incomplete occlusion on VE images demonstrated regrowth in follow-up examinations. Residual sites and volumetric ratios were correlated with aneurysmal regrowth. CONCLUSION: VE imaging can demonstrate a residual lumen more frequently than can 2D DSA imaging and is useful for evaluating aneurysmal occlusion after coil embolization.

Adult↗

Keypoint recognition using randomized trees.

In many 3D object-detection and pose-estimation problems, runtime performance is of critical importance. However, there usually is time to train the system, which we will show to be very useful. Assuming that several registered images of the target object are available, we developed a keypoint-based approach that is effective in this context by formulating wide-baseline matching of keypoints extracted from the input images to those found in the model images as a classification problem. This shifts much of the computational burden to a training phase, without sacrificing recognition performance. As a result, the resulting algorithm is robust, accurate, and fast-enough for frame-rate performance. This reduction in runtime computational complexity is our first contribution. Our second contribution is to show that, in this context, a simple and fast keypoint detector suffices to support detection and tracking even under large perspective and scale variations. While earlier methods require a detector that can be expected to produce very repeatable results, in general, which usually is very time-consuming, we simply find the most repeatable object keypoints for the specific target object during the training phase. We have incorporated these ideas into a real-time system that detects planar, nonplanar, and deformable objects. It then estimates the pose of the rigid ones and the deformations of the others.

Algorithms↗

Improvement of computerized mass detection on mammograms: fusion of two-view information.

Recent clinical studies have proved that computer-aided diagnosis (CAD) systems are helpful for improving lesion detection by radiologists in mammography. However, these systems would be more useful if the false-positive rate is reduced. Current CAD systems generally detect and characterize suspicious abnormal structures in individual mammographic images. Clinical experiences by radiologists indicate that screening with two mammographic views improves the detection accuracy of abnormalities in the breast. It is expected that the fusion of information from different mammographic views will improve the performance of CAD systems. We are developing a two-view matching method that utilizes the geometric locations, and morphological and textural features to correlate objects detected in two different views using a prescreening program. First, a geometrical model is used to predict the search region for an object in a second view from its location in the first view. The distance between the object and the nipple is used to define the search area. After pairing the objects in two views, textural and morphological characteristics of the paired objects are merged and similarity measures are defined. Linear discriminant analysis is then employed to classify each object pair as a true or false mass pair. The resulting object correspondence score is combined with its one-view detection score using a fusion scheme. The fusion information was found to improve the lesion detectability and reduce the number of FPs. In a preliminary study, we used a data set of 169 pairs of cranio-caudal (CC) and mediolateral oblique (MLO) view mammograms. For the detection of malignant masses on current mammograms, the film-based detection sensitivity was found to improve from 62% with a one-view detection scheme to 73% with the new two-view scheme, at a false-positive rate of 1 FP/image. The corresponding cased-based detection sensitivity improved from 77% to 91%.

Algorithms↗