Search

High demand of optical inspection is increased to guarantee manufacture and product quality in industries. To overcome limitations of the manual defect inspection, machine vision inspection is needed to efficiently and accurately screen the undesired defects on various products. Recently, the transparent substrate is becoming widely used for manufacturing optics and electronics products. For high-grade transparent substrates, development of machine vision inspection has increased its importance for inspecting defects after production. To perform machine vision inspection for the transparent substrate, the exposure procedure and analysis of the capturing image are critical challenges due to its properties of reflection and transparency. However, conventional machine vision systems are performed for optical inspection based on two-dimensional (2D) intensity images from the camera-based photography without phase and depth information, and may decrease inspection accuracy as well as defect classification. Conversely, instead of the 2D intensity image by camera-based photography with complicated algorithms and time-consuming computation, digital holography is a novel three-dimensional (3D) imaging technique to rapidly access the whole wavefront information of the target sample for optical inspection and complex defect analysis. In this study, we propose digital holographic imaging of transparent target sample for optical inspection in learning-based pattern classification, which a novel complex defect inspection model is presented for multiple defects identification of the transparent substrate based on 3D diffraction characteristics and machine learning algorithm. Both theoretical and experimental results will be presented and analyzed to verify the effective inspection and high accuracy.

This study proposes a novel data augmentation method based on numerical focusing of digital holography to boost the performance of learning-based pattern classification. To conduct digital holographic data augmentation (DHDA), a complex pattern diffraction approach is used to provide the least separation of confusion in the effective diffraction regime to access the full-field wavefront information of a target sample. By using DHDA, the accessible amount of labeled data is increased to complement the data manifold and to provide various three-dimensional diffraction characteristics for improving the performance of learning-based pattern classification. Experimental results demonstrated that overall accuracy of pattern classification with DHDA (95.1%) was higher than that without DHDA (90.9%).

Purpose: Breast ultrasound(US) is recently becoming more and more popular for detecting breast lesions. However, screening results in hundreds of USimages for each subject. This magnitude of images can lead to fatigue in radiologist, causing failure in the detection of lesions of a subtle nature. In this study, an image stitching technique is proposed for combining multipass images of the whole breast into a series of full-view images, and a fully automatic screening system that works off these images is also presented. Methods: Using the registration technique based on the simple sum of absolute block-mean difference (SBMD) measure, three-pass images were merged into full-view USimages. An automatic screening system was then developed for detecting tumors from these full-view images. The preprocessing step was used to reduce the tumor detection time of the system and to improve image quality. The gray-level slicing method was then used to divide images into numerous regions. Finally, seven computerized features—darkness, uniformity, width-height ratio, area size, nonpersistence, coronal area size, and region continuity—were defined and used to determine whether or not each region was a part of a tumor. Results: In the experiment, there was a total of 25 experimental cases with 26 lesions, and each case was composed of 252 images (three passes, 84 images/pass). The processing time of the proposed stitching procedure for each case was within 30 s with a Pentium IV 2.0 processor, and the detection sensitivity of the proposed CADsystem was 92.3% with 1.76 false positives per case. Conclusions: The proposed automatic screening system can be applied to the whole breast images stitched together via SBMD-based registration in order to detect tumors.

Breast density has been established as an independent risk factor associated with the development of breast cancer. The terms mammographic density and breast density are often used interchangeably, since most breast density studies are performed with projection mammography. It is known that increase in mammographic density is associated with an increased cancer risk. A sensitive method that allows for the measurement of small changes in breast density may provide useful information for risk management. Despite the efforts to develop quantitative breast density measurements from projection mammograms, the measurements show large variability as a result of projection imaging, differing body position, differing levels of compression, and variation of the x-ray beam characteristics. This study used two separate computer-aided methods, threshold-based and proportion-based evaluations, to analyze breast density on whole breast ultrasound(US)imaging and to compare with the grading results of three radiologists using projection mammography. Thirty-two female subjects with 252 images per case were included in this study. Whole breast USimages were obtained from an Aloka SSD-5500 ultrasound machine with an ASU-1004 transducer (Aloka, Japan). Before analyzing breast density, an adaptive speckle reduction filter was used for removing speckle noise, and a robust thresholding algorithm was used to divide breast tissue into fatty or fibroglandular classifications. Then, the proposed approaches were applied for analysis. In the threshold-based method, a statistical model was employed to determine whether each pixel in the breast region belonged to fibroglandular or fatty tissue. The proportion-based method was based on three-dimensional information to calculate the volumetric proportion of fibroglandular tissue to the total breast tissue. The experimental cases were graded by the proposed analysis methods and compared with the ground standard density classification assigned by a majority voting of three experienced breast radiologists. For the threshold-based method, 28 of 32 US test cases and for the proportion-based density classifier, 27 of 32 US test cases were found to be in agreement with the radiologist “ground standard” mammographic interpretations, resulting in overall accuracies of 87.5% and 84.4%, respectively. Moreover, the concordance values of the proposed methods were between 0.0938 and 0.1563, which were less than the average interobserver concordance of 0.3958. The experiment result showed that the proposed methods could be a reference opinion and offer concordant and reliable quantification of breast density for the radiologist.

Ultrasound imaging plays an important role in the field of breast cancer diagnosis because of its convenience and non-invasive. Recently, the development of computer-aided diagnosis (CAD) provides a convenient way for doctors in detecting breast cancer using ultrasound images. However, the previous CAD systems have some limits with the requirements of human intervention. Hence, in this paper, a novel automatic CAD system is proposed to find suspicious frames among whole breast ultrasound images. After applying watershed segmentation, suspicious segmented regions can be identified through several criteria defined according to the statistic and geometric features of a tumour. By examining 13 US test cases, almost all the tumours and cysts could successfully be detected when an average of two false positives for each case is produced. The experimental results prove the accuracy of this proposed system.

Breast ultrasound (US) is recently becoming more and more popular for detecting breast lesions. However, screening results in hundreds of US images for each subject. This magnitude of images can lead to fatigue in radiologist, causing failure in the detection of lesions of a subtle nature. In this study, an image stitching technique is proposed for combining multipass images of the whole breast into a series of full-view images, and a fully automatic screening system that works off these images is also presented.Using the registration technique based on the simple sum of absolute block-mean difference (SBMD) measure, three-pass images were merged into full-view US images. An automatic screening system was then developed for detecting tumors from these full-view images. The preprocessing step was used to reduce the tumor detection time of the system and to improve image quality. The gray-level slicing method was then used to divide images into numerous regions. Finally, seven computerized features--darkness, uniformity, width-height ratio, area size, nonpersistence, coronal area size, and region continuity--were defined and used to determine whether or not each region was a part of a tumor.In the experiment, there was a total of 25 experimental cases with 26 lesions, and each case was composed of 252 images (three passes, 84 images/pass). The processing time of the proposed stitching procedure for each case was within 30 s with a Pentium IV 2.0 processor, and the detection sensitivity of the proposed CAD system was 92.3% with 1.76 false positives per case.The proposed automatic screening system can be applied to the whole breast images stitched together via SBMD-based registration in order to detect tumors.

Breast density has been established as an independent risk factor associated with the development of breast cancer. The terms mammographic density and breast density are often used interchangeably, since most breast density studies are performed with projection mammography. It is known that increase in mammographic density is associated with an increased cancer risk. A sensitive method that allows for the measurement of small changes in breast density may provide useful information for risk management. Despite the efforts to develop quantitative breast density measurements from projection mammograms, the measurements show large variability as a result of projection imaging, differing body position, differing levels of compression, and variation of the x-ray beam characteristics. This study used two separate computer-aided methods, threshold-based and proportion-based evaluations, to analyze breast density on whole breast ultrasound (US) imaging and to compare with the grading results of three radiologists using projection mammography. Thirty-two female subjects with 252 images per case were included in this study. Whole breast US images were obtained from an Aloka SSD-5500 ultrasound machine with an ASU-1004 transducer (Aloka, Japan). Before analyzing breast density, an adaptive speckle reduction filter was used for removing speckle noise, and a robust thresholding algorithm was used to divide breast tissue into fatty or fibroglandular classifications. Then, the proposed approaches were applied for analysis. In the threshold-based method, a statistical model was employed to determine whether each pixel in the breast region belonged to fibroglandular or fatty tissue. The proportion-based method was based on three-dimensional information to calculate the volumetric proportion of fibroglandular tissue to the total breast tissue. The experimental cases were graded by the proposed analysis methods and compared with the ground standard density classification assigned by a majority voting of three experienced breast radiologists. For the threshold-based method, 28 of 32 US test cases and for the proportion-based density classifier, 27 of 32 US test cases were found to be in agreement with the radiologist "ground standard" mammographic interpretations, resulting in overall accuracies of 87.5% and 84.4%, respectively. Moreover, the concordance values of the proposed methods were between 0.0938 and 0.1563, which were less than the average interobserver concordance of 0.3958. The experiment result showed that the proposed methods could be a reference opinion and offer concordant and reliable quantification of breast density for the radiologist.

Characterization and quantification of diffuse parenchymal lung disease (DPLD) severity using MDCT, mainly in interstitial lung diseases and emphysema, is an important issue in clinical research for the evaluation of new therapies. This paper develops a 3D automated approach for detection and diagnosis of DPLDs (emphysema, fibrosis, honeycombing, ground glass). The proposed methodology combines multi-resolution image decomposition based on 3D morphological filtering, and graph-based classification for a full characterization of the parenchymal tissue. The very promising results obtained on a small patient database are good premises for a near implementation and validation of the proposed approach in clinical routine.

Characterization and quantification of diffuse parenchymal lung disease (DPLD) severity using MDCT, mainly in interstitial lung diseases and emphysema, is an important issue in clinical research for the evaluation of new therapies. This paper develops a 3D automated approach for detection and diagnosis of DPLDs (emphysema, fibrosis, honeycombing, ground glass).The proposed methodology combines multiresolution image decomposition based on 3D morphological filtering, and graph-based classification for a full characterization of the parenchymal tissue. The very promising results obtained on a small patient database are good premises for a near implementation and validation of the proposed approach in clinical routine.

In general, several factors are used for risk estimation in breast cancer detection and early prevention, and one of the important factors in risk of breast cancer is breast density. The mammography is important and effective adjunct in diagnosing the breast cancer. The radiologists would analyze visually the breast density with the BI-RADS lexicon on mammograms. However, this usually causes a large inter-observer variability among the different experienced radiologists. In this paper, we individually adopt three methods, including pixel-based, region-based, and physics-based, to analyze the breast density on mammograms, and the results can offer radiologists a second quantification reading for predicting the risk of breast cancer. The three methods are tested on 208 digital and conventional film mammograms which are scanned from both breasts of 104 patients respectively. The experimental results show that the accuracy of the proposed region-based method, which is more consistent with the radiologists' viewpoint, is 88% more than other two conventional methods.

Diffuse lung diseases (DLD) include a heterogeneous group of non-neoplasic disease resulting from damage to the lung parenchyma by varying patterns of inflammation. Characterization and quantification of DLD severity using MDCT, mainly in interstitial lung diseases and emphysema, is an important issue in clinical research for the evaluation of new therapies. This paper develops a 3D automated approach for detection and diagnosis of diffuse lung diseases such as fibrosis/honeycombing, ground glass and emphysema. The proposed methodology combines multi-resolution 3D morphological filtering (exploiting the sup-constrained connection cost operator) and graph-based classification for a full characterization of the parenchymal tissue. The morphological filtering performs a multi-level segmentation of the low- and medium-attenuated lung regions as well as their classification with respect to a granularity criterion (multi-resolution analysis). The original intensity range of the CT data volume is thus reduced in the segmented data to a number of levels equal to the resolution depth used (generally ten levels). The specificity of such morphological filtering is to extract tissue patterns locally contrasting with their neighborhood and of size inferior to the resolution depth, while preserving their original shape. A multi-valued hierarchical graph describing the segmentation result is built-up according to the resolution level and the adjacency of the different segmented components. The graph nodes are then enriched with the textural information carried out by their associated components. A graph analysis-reorganization based on the nodes attributes delivers the final classification of the lung parenchyma in normal and ILD/emphysematous regions. It also makes possible to discriminate between different types, or development stages, among the same class of diseases.

Over the past few years, the billows of the digital trends and the exploding growth of electronic networks, such as worldwide web, global mobility networks, etc., have drastically changed our daily lifestyle. In view of the widespread applications of digital images, medical images, which are produced by a wide variety of medical appliances, are stored in digital form gradually. These digital images are very easy to be modified imperceptively by malicious intruders for illegal purposes. The well-known adage that “seeing is believing” seems not always a changeless truth. Therefore, protecting images from being altered becomes an important issue. Based on the lossless data-embedding techniques, two detection and restoration systems are proposed to cope with forgery of medical images in this paper. One of them has the ability to recover the whole blocks of the image and the other enables to recover only a particular region where a physician will be interested in, with a better visual quality. Without the need of comparing with the original image, these systems have a great advantage of detecting and locating forged parts of the image with high possibility. And then it can also restore the counterfeited parts. Furthermore, once an image is announced authentic, the original image can be derived from the stego-image losslessly. The experimental results show that the restored version of a tampered image in the first method is extremely close to the original one. As to the second method, the region of interest selected by a physician can be recovered without any loss, when it is tampered.

The breast density information is one of important factors for estimating the risk in breast cancer detection and early prevention. In this paper, we present two methods, including threshold-based and proportion-based, to automatically analyze the breast density using whole breast ultrasound. The two algorithms are experimented with 32 cases which are scanned from 32 patients using the US machine SSD-5500 with a recent developed scanner ASU-1004 (Aloka, Japan). The experimental results are graded from 4 (extremely dense tissue) to 1 (almost entirely fat), and respectively compared with the majority grades of three radiologists. The accuracy of the threshold-based and proportion-based strategies is 88% and 84% respectively.