AUTOMATED VESSEL SEGMENTATION FROM RETINA USING MATLAB

AUTOMATED VESSEL SEGMENTATION FROM RETINA 

 

          Blood vessels can be conceptualized anatomically as an intricate network, or tree-like structure (or vasculature), of hollow tubes of different sizes and compositions including arteries, arterioles, capillaries, venules, and veins. Their continuing integrity is vital to nurture life: any damage to them could lead to profound complications, including stroke, diabetes, arteriosclerosis, cardiovascular diseases and hypertension, to name only the most obvious. Vascular diseases are often life-critical for individuals, and present a challenging public health problem for society. The drive for better understanding and management of these conditions naturally motivates the need for improved imaging techniques. The detection and analysis of the vessels in medical images is a fundamental task in many clinical applications to support early detection, diagnosis and optimal treatment.

Fig: (A) A randomly chosen image from the DRIVE dataset. (B)-(D) Enhancement results on (A) by using the eigenvalue-based (FR), wavelet-based (IUWT), and local phase-based (LP) filters respectively. (E) Expert’s annotation.

In line with the proliferation of imaging modalities, there is an ever-increasing demand for automated vessel analysis systems for which where blood vessel segmentation is the first and most important step. As blood vessels can be seen as linear structures distributed at different orientations and scales in an image, various kernels (or enhancement filters) have been proposed to enhance them in order to ease the segmentation problem. In particular, a local phase based filter recently introduced by Lathen et al seems to be superior to intensity based filters as it is immune to intensity inhomogeneity and is capable of faithfully enhancing vessels of different widths.

FOR MORE DETAILS 

                                                      CLICK HERE

SEGMENTATION OF PHEOCHROMOCYTOMAS USING MATLAB?

SEGMENTATION OF PHEOCHROMOCYTOMAS IN CECT IMAGES

          Segmentation of pheochromocytomas in Contrast-Enhanced Computed Tomography (CECT) images is an ill-posed problem due to the presence of weak boundaries, intratumoral degeneration, and nearby structures and clutter.

            Simultaneously co-segmenting common objects from a pair of images has drawn much attention. In such cases, the region-based LSMs (RLSMs) are more suitable by using statistical information of foreground and background regions. To improve the capability of segmenting objects having heterogeneous regions, local image information are widely considered in many localized RLSMs (LRLSMs).

Fig:REGION OF TUMOR

FOR DETAILS ABOUT LATEST IEEE PROJECTS

          Adaptive Co-Segmentation of Pheochromocytomas in CECT Images Using Localized Level Set Models

DETECTION OF TUMOR AND THROMBI USING MATLAB?

DETECTION OF TUMOR AND THROMBI

Intracardiac masses are hazardous in cardiovascular disease. Generally, they are abnormal structures within or immediately adjacent to the heart, which must be distinguished for diagnosis. Two main types of intracardiac masses are tumor and thrombus. Identification of intracardiac masses in echocardiograms is one important task in cardiac disease diagnosis. To improve diagnosis accuracy, a novel fully automatic classification method based on the sparse representation is proposed to distinguish intracardiac tumor and thrombi in echocardiography.

The tumors show continuity with the atrial wall, with a high degree of mobility, as shown. The echocardiographic appearances of the thrombi are motionless, dense, ovoid, and echo reflecting mass with a broad base of attachment to the endocardium.

                                                   Fig: DETECTION OF TUMOR

adaptive cancer detction using MATLAB?

FAST AND ADAPTIVE DETECTION OF PULMONARY NODULES IN THORACIC CT IMAGES

PULMONARY NODULES

In general, a “pulmonary nodule” is a small, roundish growth on the lung that measures three centimeters in diameter or less. If the growth is larger than that, it is called a “pulmonary mass.” While pulmonary nodules may grow to become a pulmonary mass, some nodules may not grow at all. There are many causes of pulmonary nodules. These include infection, such as fungal or bacterial infections, noncancerous processes, such as sarcoidosis, or cancerous processes, such as lung cancer, lymphoma, or metastatic cancer from other organs. The likelihood that a pulmonary nodule represents lung cancer depends upon three major factors, your age, your smoking history, and your environmental exposure history. Generally, less than 10 percent of pulmonary nodules turn out to be lung cancer.

                                                 Figure.  CT scan image with a Pulmonary nodule

SYMPTOMS OF PULMONARY NODULES

Because pulmonary nodules are small, they rarely cause any symptoms. Some patients might experience symptoms of a respiratory infection, such as a symptoms associated with chest colds or mild flu. Most pulmonary nodules are discovered by accident, when a patient gets a chest X-ray or a CT scan performed for another purpose. 

EVALUATING A PULMONARY NODULE

The immediate goal of evaluating a pulmonary nodule is determining the cancerous potential of the nodule. This is first done with a thorough evaluation of the personal and medical history, the environmental exposure history, and the chest CT scan. If a nodule is determined to have significant cancer potential and is one centimeter in diameter or greater, diagnostic procedures are used to determine the cause of the pulmonary nodule. There are many approaches to evaluating and diagnosing pulmonary nodules that do not require surgery, such as PET scans, bronchoscopy, endobronchial ultrasound, CT-guided needle biopsy, and fluoroscopically guided biopsy. When the pulmonary nodule cannot be diagnosed using these noninvasive approaches, surgical approaches are considered, such as video-assisted thorocoscopic surgery, a mini-thoracotomy, or a thoracotomy. Once the cause of the pulmonary nodule has been determined, an appropriate treatment plan tailored to the disease can be assembled.

FOLLOWING A PULMONARY NODULE

       The majority of pulmonary nodules are extremely small, less than one centimeter in diameter. Unfortunately, these pulmonary nodules are too small to be diagnosed safely and accurately using any of the currently available procedures or tests. Because these very small pulmonary nodules can represent early lung cancer, they need to be followed closely using CT scans with a well developed algorithm for evaluating whether the pulmonary nodule has grown over time. If the size of these pulmonary nodules remains unchanged for two years, the likelihood of these pulmonary nodules representing lung cancer is very small.

 FOR MORE DETAILS

                           click here

FAST AND ADAPTIVE DETECTION OF PULMONARY NODULES

FAST AND ADAPTIVE DETECTION OF PULMONARY NODULES IN THORACIC CT IMAGES

PULMONARY NODULES

In general, a “pulmonary nodule” is a small, roundish growth on the lung that measures three centimeters in diameter or less. If the growth is larger than that, it is called a “pulmonary mass.” While pulmonary nodules may grow to become a pulmonary mass, some nodules may not grow at all. There are many causes of pulmonary nodules. These include infection, such as fungal or bacterial infections, noncancerous processes, such as sarcoidosis, or cancerous processes, such as lung cancer, lymphoma, or metastatic cancer from other organs. The likelihood that a pulmonary nodule represents lung cancer depends upon three major factors, your age, your smoking history, and your environmental exposure history. Generally, less than 10 percent of pulmonary nodules turn out to be lung cancer.

                                                 Figure.  CT scan image with a Pulmonary nodule

SYMPTOMS OF PULMONARY NODULES

Because pulmonary nodules are small, they rarely cause any symptoms. Some patients might experience symptoms of a respiratory infection, such as a symptoms associated with chest colds or mild flu. Most pulmonary nodules are discovered by accident, when a patient gets a chest X-ray or a CT scan performed for another purpose. 

EVALUATING A PULMONARY NODULE

The immediate goal of evaluating a pulmonary nodule is determining the cancerous potential of the nodule. This is first done with a thorough evaluation of the personal and medical history, the environmental exposure history, and the chest CT scan. If a nodule is determined to have significant cancer potential and is one centimeter in diameter or greater, diagnostic procedures are used to determine the cause of the pulmonary nodule. There are many approaches to evaluating and diagnosing pulmonary nodules that do not require surgery, such as PET scans, bronchoscopy, endobronchial ultrasound, CT-guided needle biopsy, and fluoroscopically guided biopsy. When the pulmonary nodule cannot be diagnosed using these noninvasive approaches, surgical approaches are considered, such as video-assisted thorocoscopic surgery, a mini-thoracotomy, or a thoracotomy. Once the cause of the pulmonary nodule has been determined, an appropriate treatment plan tailored to the disease can be assembled.

FOLLOWING A PULMONARY NODULE

       The majority of pulmonary nodules are extremely small, less than one centimeter in diameter. Unfortunately, these pulmonary nodules are too small to be diagnosed safely and accurately using any of the currently available procedures or tests. Because these very small pulmonary nodules can represent early lung cancer, they need to be followed closely using CT scans with a well developed algorithm for evaluating whether the pulmonary nodule has grown over time. If the size of these pulmonary nodules remains unchanged for two years, the likelihood of these pulmonary nodules representing lung cancer is very small.

 FOR MORE DETAILS

                           click here

DETECTION OF TUMOR

DETECTION OF TUMOR AND THROMBI

Intracardiac masses are hazardous in cardiovascular disease. Generally, they are abnormal structures within or immediately adjacent to the heart, which must be distinguished for diagnosis. Two main types of intracardiac masses are tumor and thrombus. Identification of intracardiac masses in echocardiograms is one important task in cardiac disease diagnosis. To improve diagnosis accuracy, a novel fully automatic classification method based on the sparse representation is proposed to distinguish intracardiac tumor and thrombi in echocardiography.

The tumors show continuity with the atrial wall, with a high degree of mobility, as shown. The echocardiographic appearances of the thrombi are motionless, dense, ovoid, and echo reflecting mass with a broad base of attachment to the endocardium.

                                                   Fig: DETECTION OF TUMOR

FOR LATEST IEEE PAPER 

Automatic Classification of Intracardiac Tumor and Thrombi in Echocardiography Based on Sparse Representation

FOUR-CLASS CLASSIFICATION OF SKIN LESIONS WITH TASK DECOMPOSITION STRATEGY

CLASSIFICATION OF SKIN LESIONS

           Incidence of skin cancer has been increasing over the decades and early treatment is becoming more and more important. The five year survival rate of melanoma, the most fatal skin cancer is only 9–15% at stage IV, while this rate increases to 85–99% if detected early at stage II. Basal cell carcinoma (BCC), the most common skin cancer is rarely fatal, but it destroys surrounding tissue if left untreated. 

Fig: Four regions in the skin lesion image.

In this paper, we focus on the first issue, i.e., the limitation of applicable skin lesion types. That is, most of the conventional works handled only melanocytic skin lesions (MSLs) such as melanomas and nevi, which originate from melanocytes, whereas nonmelanocytic skin lesions, (NoMSLs) indicating all the other pigmented skin lesions except MSLs such as BCCs and seborrheic keratoses (SKs) have been relatively neglected.

FOR MORE DETAILS                           

                                                  CLICK HERE