Marek Pazurek

Texture and color based image segmentation and pathology detection in capsule endoscopy videos

This paper presents an in-depth study of several approaches to exploratory analysis of wireless capsule endoscopy images (WCE). It is demonstrated that versatile texture and color based descriptors of image regions corresponding to various anomalies of the gastrointestinal tract allows their accurate detection of pathologies in a sequence of WCE frames. Moreover, through classification of single pixels described by texture features of their neighborhood, the images can be segmented into homogeneous areas well matched to the image content. For both, detection and segmentation tasks the same procedure is applied which consists of features calculation, relevant feature subset selection and classification stages. This general three-stage framework is realized using various recognition strategies. In particular, the performance of the developed Vector Supported Convex Hull classification algorithm is compared against Support Vector Machines run in configuration with two different feature selection methods.

Pancreatic cyst fluid analysis for differential diagnosis between benign and malignant lesions

The majority of pancreatic cysts are detected incidentally when abdominal imaging is performed during unrelated procedures. The aim of the present study was to assess the diagnostic utility and clinical value of carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA 19-9) and amylase analysis in pancreatic cyst fluid. The study included 52 patients with pancreatic cystic lesions, who underwent fine-needle aspiration biopsy to collect cystic fluid for cytological and biochemical analysis. Cysts were classified as benign (simple cysts, pseudocysts and serous cystadenomas) in 36 patients or premalignant/malignant (mucinous cyst-adenomas, intraductal papillary mucinous neoplasm and cystadenocarcinomas) in 16 patients. CEA and CA 19-9 were elevated in patients with malignant cysts (238±12.5 ng/ml and 222±31.5 U/ml, respectively) compared with benign lesions (34.5±3.7 ng/ml and 18.5±1.9 U/ml, respectively; P<0.001). Based on these results, the sensitivity and specificity of CEA were 91.8 and 63.9% and of CA 19-9 were 81.3 and 69.4%, respectively. Mean amylase levels in benign lesions (27825.7±91.9 U/l) were higher compared with malignant pancreatic cysts (8359.2±32.7 U/l; P<0.05). Cyst fluid analysis may prove a safe and useful adjunct for the differential diagnosis of pancreatic cystic lesions. In the present study, promising results for CEA and CA 19-9 have been demonstrated, however, the clinical value of these molecules must be confirmed.

Photodynamic therapy in the palliative treatment of esophageal cancer

Esophageal carcinoma is one of the most common cancers in the world and has an increasing incidence in Western civilisation and poor prognosis. In most cases palliative treatment is the only kind of therapy, which can be applied. In the Department of Digestive Tract Diseases, Medical University, Lodz, Poland, we have used PDT in two cases of palliative therapy of advanced esophageal cancer in cervical location of esophagus. We find PDT very promising, minimally invasive, safe and easy technique to perform for esophageal cancer of cervical localization.

Local polynomial approximation for unsupervised segmentation of endoscopic images

In this paper we present a novel technique for unsupervised texture segmentation of wireless capsule endoscopic images of the human gastrointestinal tract. Our approach integrates local polynomial approximation algorithm with the well-founded methods of color texture analysis and clustering (k-means) leading to a robust segmentation procedure which produces fine-grained segments well matched to the image contents.