Mircea-Florin Vaida

Medical image fusion based on discrete wavelet transform using Java technology

The importance of information offered by the medical images for diagnosis support can be increased by combining images from different compatible medical devices. The fusion process allows combination of salient feature of these images. In this paper we present different techniques of image fusion, our work for medical image fusion based on discrete wavelet transform and how we understand to integrate this process into a distributed application. The dedicated application considers Java technology for using its facilities as a future development, regarding a remote access mechanism.

SOA-based medical image processing platform

The aim of this paper is to present the design of a service oriented architecture-based platform for medical image processing in assisted diagnosis. Service oriented architecture (SOA) improves the reusability and maintainability of distributed systems. In service oriented architectures, the most important element is the service, a resource provided to remote clients via a service contract. SOA-based systems may prove to be a good solution to address problems like large storage and expensive computational requirements faced in medical image processing applications. We propose a component-based platform which is not tied to a specific programming language or a specific technology. Our platform provides a repository that contains components used to create services.

Elastic-net based beamforming in medical ultrasound imaging

This paper presents a new way of addressing beamforming in ultrasound imaging, by formulating it, for each image depth, as an inverse problem solved using elastic-net regularization. This approach was evaluated on both simulated and in vivo data showing a gain in contrast, while maintaining an increased value of the signal-to-noise ratio compared to two standard ultrasound beamforming methods.

Beamforming with sparse prior in ultrasound medical imaging

Nowadays the classical Delay-and-Sum (DAS) beamformer is extensively used in ultrasound imaging due to its low computational characteristics. However, it suffers from high sidelobe level, poor resolution and low contrast. An alternative is the Minimum-Variance (MV) beamformer which results in a higher image quality both in terms of spatial resolution and contrast. Even so, these benefits come at the expense of a higher computation complexity that limits its real-time capabilities. One solution that recently gained noticeable interest is the exploit of the sparsity of the scanned medium. Based on this assumption, we extend the DAS method to yield sparse results by using the Bayesian Information Criterion (BIC). Our realistic simulations demonstrate that the proposed beamforming (BF) method shows better performance than the classical DAS and MV in terms of lateral resolution, sidelobe reduction and contrast.

On-demand service composition based on natural language requests

The widespread and diversity of Web services in the ubiquitous computing era and the impossibility to predict a priori all possible user demands, generates the need of a system capable of dynamically composing new services, based on unrestricted natural language requests.

Service oriented architecture for medical image processing

ObjectiveOne of the problem existent today in the medical domain is that we have an important number of legacy applications that cannot be reused, shared, composed. The objective of this paper is to propose a solution that make possible to overcome these limitations and makes possible the creation of new composite services that may be used in common by several institutions.MethodsA generic model for a service, based on a loosely coupled, message-based communication model is proposed. The service model takes into account the possibility to integrate legacy applications. Specialized image processing services can be dynamically discovered and integrated into client applications or other services. A new model and software product based on Service Oriented Architecture (SOA) is proposed. SOA improves the reusability and maintainability of distributed systems. In service oriented architectures, the most important element is the service, a resource provided to remote clients via a service contract.ResultsA basic integration example is presented in the paper and the possibility to integrate dedicated scenario.ConclusionThe aim of this paper is to present a services based architecture for medical image processing in assisted diagnosis. A service can be discovered at runtime, can be replaced if it is unavailable or can be incorporated in a new service.

Enhancing the security of Web applications

The main aim is to present the security elements involved in Web applications. E-commerce applications and data access applications such as a university information database may be protected using different security techniques. Considering an e-commerce application, we developed some security elements that will provide an accurate and safe data transfer mechanism. We adapted the application for allowing our university students to access their exam results and to view their overall situation or some private data using Web facilities.

Towards Real-Life Facial Expression Recognition Systems

Facial expressions are a set of symbols of great importance for human-to-human communication. Spontaneous in their nature, diverse and personal, facial expressions demand for real-time ...

Active contours: Application to plant recognition

The problem we address in this paper is object segmentation applied to plant recognition. The image can contain one or more plants on a natural background. More precisely, we aim to segment flowers. This approach poses several challenges, such as texture, multiple colors that form one object, natural background, non-homogeneous regions, etc. We propose an approach that adapts the Chan-Vese model, [1], in order to use it with a fast level-set approximation algorithm, [2]. Our approach presents ongoing work towards flower segmentation and recognition. In this paper we present the main outcome of our research on segmentation problem (curve evolution from Chan-Vese model with a recent and faster approach, [2], and its optimization at implementation level) and on recognition problem (identifying flower species). At recognition level, we studied five features that may be extracted during segmentation. Experiments have been carried out over the Oxford Flowers dataset, [3].

A Web 3.0 solution for restraining the web-bots access to the on-line displayed content

This paper presents a technique meant for eliminating the access possibilities that web-bots can use in order to perform automatically some on-line tasks destined to the human operators. The applications that benefit from the proposed solution are client-to-business or business-to-business web distributed products. The solution refers to protecting the e-forms embedded into the web pages and the implementation can be regarded as a CAPTCHA. Unlike the classical approach to the matter, the solution inserts no supplementary fields in the pages content. The modules loaded onto the client machines can make the difference between a human person and a web robot.