Hulya Dogan

Image panaroma with autofocusing on microscopic imaging

During microscopic analyses for recognition of bacteria, the prior process is to provide the optimum focusing. After the realization of focusing, due to small field o f v iew on the microscope microbiology laboratory technicians must scan without losing focusing by controlling with eye-hand to see all part of sample. This process takes a lot of time and increases making mistakes during laboratory analyses. In this study, minimizing these negative effects and obtaining more reliable results in the bacteria analysis process are proposed. Firstly passive focus methods based autofocusing on the Z axis of microscope was implemented. After that, automatic scanning begins by moving microscope platform with step motors on the X-Y axes. During scanning process, determining the focusing degradation images on the X-Y axes was achieved, high resolution (panoramic) image was obtained with panorama without losing focusing. Moreover, panorama was generated with and without losing focusing and it was shown that the image obtained without losing focusing had better quality and higher resolution. Proposed study was tested on image sequence with smear stained sputum for TB diagnosis

Microscopic image segmentation based on firefly algorithm for detection of tuberculosis bacteria

One third of the world is infected with tuberculosis disease. The disease is diagnosed visually by laboratory technicians. In the microscopy diagnosis with hand-eye control, misdiagnosis rate is quite high. In microscopic imaging, by using computer aided automatic diagnosis methods, the disease is true diagnosed. The robustness of the automatic diagnosis methods depends on accurate segmentation of microscopic images. Image segmentation methods produce a special solution for several problems. In this study, Firefly algorithm based on swarm intelligence as a novel approach in microscopic imaging is proposed to segment images. In the proposed approach, an optimum threshold value in gray-level microscopic images is determined with proposed entropy based Firefly algorithm. Microscopic images are converted to binary format by using obtained optimum threshold value. Segmentation results are compared with expert-guided segmentation results. The performance ratio of segmentation is 96% obtained by using Firefly algorithm based on swarm intelligence.

Image forgery detection based on Colour SIFT

Nowadays several modifications on the digital images are made with the rapid development of image editing tools in recent years. The most common method in the modifications made is copy-move forgery. Majority of the proposed methods to detect copy-move forgery in the literature are based on block and not resistant to various geometric transformation before moving coped image. For this aim the key points of each channel of forged color image are extracted by using Colour SIFT that is the key point-based method. In this study the comparison between SIFT and Colour SIFT is made. More effective forgery detection is made by obtaining more matching points using Colour SIFT than SIFT is seen in the comparison results. Moreover, proposed method detects forged image during rotation, scaling, JPEG compression is shown in the results.

Auto-focusing with multi focus color image fusion based on curvelet transform on microscopic imaging

The fundamental operation before analyzing bacteria on the microscopic system is optimal focusing. Laboratory technicians implement this process with eye-hand coordination. During auto-focus process avoiding the dependence on technicians, auto-focus functions giving a value about focusing of the images are used in literature. At the end of the auto-focusing based on auto-focus functions, some regions in the in-focus image can be blurred. In this paper auto-focusing based color image fusion is implemented to obtain all of region in-focus image. In this study, firstly an image sequence is captured with moving the microscope stage along Z-axis. The reference image with the highest focus value on the sequence is found. Reference image and several images around the reference image are fused with curvelet transform preferred to obtain curve and line information of image. Moreover, various evaluation criteria are utilized to analyze the performance of the proposed auto-focus approach on different color models.

Auto-focusing on microscopic imaging with image fusion method

The priority operation on process of analyzing bacteria automatically on the microscopic images is realization of optimal focusing. Reliable imaging can be achieved by examining multiple images in sequential images formed with Hand-eye coordination of the laboratory on Z axis of microscope. In this study, instead of selection of a single frame with maximum autofocus function value on sequential images, automatic generation of optimal color image based on the processing more frames around frame with maximum auto-focus value is proposed. Microscopic imaging with optimal focusing is generated by processing images with different focusing on the Z axis of microscope. Firstly, Probability density function (pdf) of image focus based on auto-focus function is obtained. The image with highest value is assigned to the reference image. Optimal focusing in all areas of color image is achieved by applying image fusion approach in a multi-dimensional color space of reference image and images with pdf values near it to ensure optimum focus in all areas(in pixels) of the image. Proposed method is compared with other methods in the literature by testing in image sequences taken from a different slides containing sputum smear painted with ARB for the diagnosis of TB.

Automated Cell Nuclei Segmentation in Pleural Effusion Cytology Using Active Appearance Model

Pleural effusion is common in clinical practice and it is a frequently encountered specimen type in cytopathological assessment. In addition to being time-consuming and subjective, this assessment also causes inter-observer and intra-observer variability, and therefore an automated system is needed. In visual examination of cytopathological images, cell nuclei present significant diagnostic value for early cancer detection and prevention. So, efficient and accurate segmentation of cell nuclei is one of the prerequisite steps for automated analysis of cytopathological images. Nuclei segmentation also yields the following automated microscopy applications, such as cell counting and classification. In this paper, we present an automated technique based on active appearance model (AAM) for cell nuclei segmentation in pleural effusion cytology images. The AAM utilizes from both the shape and texture features of the nuclei. Experimental results indicate that the proposed method separates the nuclei from background effectively. In addition, comparisons are made with the segmentation methods of thresholding-based, clustering-based and graph-based, which show that the results obtained with the AAM method are actually more closer to the ground truth.

Image panorama without loss of focusing for microscopic systems

When microbiology laboratory technicians examine the specimen on the microscope for recognition of bacteria, firstly they provide optimum focusing by adjusting of the microscope stage on the Z axis. Following the achievement of the optimal focus, the movement of the microscope stage on 3D axis (X-Y-Z) are performed without losing focusing in order to scan whole view of the specimen. Therefore, microscope diagnosis analysis consumes a lot of time and is also dependent on the experience of the technicians. In this study, instead of manual scanning of the specimen obtained by microbiology laboratory technicians, it is aimed to achieve a panoramic imaging with high resolution including whole view of the specimen in-focus by designing fully automatic motorized microscope system. In proposed study, firstly auto-focusing based on passive focus methods is performed to find an in-focus image by adjusting the position of the stage on the Z axis. After implementation of auto-focusing on the Z axis, the automatic scanning process for monitoring specimen is started by moving stage on the X-Y axes with stepper motors. During scanning process, the images having near focal length are acquired to maintain focusing; therefore, it is achieved to obtain a panoramic image higher resolution. Moreover, image stitching process is also performed with and without focusing to show that autofocusing is crucial to obtain high quality panoramic imaging. The experimental results indicate that if in focus images are used for image stitching process the more feature points are produced, and panoramic image with higher resolution more quality is obtained. To compare the performance of image stitching process with and without focusing, PSNR, RMSE and UIQI are used, and the success of the proposed study is finally proved.

Virtual mouse control with hand gesture information extraction and tracking

Virtual mouse implemented with hand gesture tracking based on image is one of studies in human-computer interaction. In this study that human-computer interaction is implemented with virtual mouse is purposed. This study consists of three main steps that are hand gesture tracking, features of hand region extraction and classification of these features. In this study hand gesture tracking is generated with Camshift (Continuously Adaptive Mean Shift) algorithm, features of hand gestures are extracted with bag of visual words and these features are classified with support vector machines. Detailed tests are performed to compare success of tracking, features extraction and classisfication methods and that this system works successfully is shown.