Название: Feature Extraction in Medical Image Retrieval: A New Design of Wavelet Filter Banks
Автор: Aswini Kumar Samantaray, Amol D. Rahulkar
Издательство: Springer
Год: 2024
Страниц: 162
Язык: английский
Формат: pdf (true), epub
Размер: 16.1 MB

Medical imaging is fundamental to modern healthcare, and its widespread use has resulted in creation of image databases. These repositories contain images from a diverse range of modalities, multidimensional as well as co-aligned multimodality images. These image collections offer opportunity for evidence-based diagnosis, teaching, and research. Advances in medical image analysis over last two decades shows there are now many algorithms and ideas available that allow to address medical image analysis tasks in commercial solutions with sufficient performance in terms of accuracy, reliability and speed. Content-based image retrieval (CBIR) is an image search technique that complements the conventional text-based retrieval of images by using visual features, such as color, texture, and shape, as search criteria. This book emphasizes the design of wavelet filter-banks as efficient and effective feature descriptors for medical image retrieval.

Firstly, a generalized novel design of a family of multiplier-free orthogonal wavelet filter-banks is presented. In this, the dyadic filter coefficients are obtained based on double-shifting orthogonality property with allowable deviation from original filter coefficients. Next, a low complex symmetric Daub-4 orthogonal wavelet filter-bank is presented. This is achieved by slightly altering the perfect reconstruction condition to make designed filter-bank symmetric and to obtain dyadic filter coefficients. In third contribution, the first dyadic Gabor wavelet filter-bank is presented based on slight alteration in orientation parameter without disturbing remaining Gabor wavelet parameters. In addition, a novel feature descriptor based on the design of adaptive Gabor wavelet filter-bank is presented. The use of Maximum likelihood estimation is suggested to measure the similarity between the feature vectors of heterogeneous medical images. The performance of the suggested methods is evaluated on three different publicly available databases namely NEMA, OASIS and EXACT09. The performance in terms of average retrieval precision, average retrieval recall and computational time are compared with well-known existing methods.

The orthogonal wavelet FBs consume large resources for hardware implementation due to their irrational coefficients. This results in reduction in speed of operation, huge memory requirement, and increased power dissipation. In order to solve these issues, a low complex symmetric db-4 orthogonal wavelet FB is presented in this book. This is achieved by slightly altering the perfect reconstruction (PR) condition to make designed FB symmetric and to obtain dyadic filter coefficients. The designed wavelet FB is multiplier free and achieves considerable reduction in hardware and dynamic power consumption by reducing number of adders and shifters. This is verified by implementing the designed wavelet FB on Kintex-7 field programmable gate array (FPGA) using vivado 2019.2 software from Xilinx. The designed FB is tested for its suitability in medical image retrieval application. Simulation results prove that the designed wavelet FB gives better performances in terms of retrieval accuracy (ARP, ARR) for medical image retrieval on the benchmark images as compared to state-of-art methods.

A computer based system used for browsing, searching, and retrieving images from a sizable collection of digital images is called a content-based image retrieval (CBIR) system. Large archives of images and multimedia content have been created in recent years due to the rapid rise of digital computers, multimedia, and storage technologies. These advancements in digital storage and content processing are also helping clinical and diagnostic studies.

Given its application in managing the gigabytes of unlabeled image data that are created and digitally stored in enormous repositories, containing visual information on the web as well as in the network computing system, automatic image retrieval has grown in importance as an area of research. These images are kept in many databases located all over the world as unstructured multimedia data, the majority of which are accessible online. The ease with which consumers may create and store images on today’s consumer electronics devices is responsible for this advancement. There are solid grounds to think that this expansion will pick up speed in the future, requiring search engine giants like Google, Bing, and others to invest a significant amount of processing power in their image search engines. Still, it is difficult to find trustworthy textual information about these images, which makes developing efficient imagine retrieval algorithms much more difficult. Furthermore, consistency can be a significant issue in manually labeled collections because a human annotator might fail to include a pertinent caption for an image or disagree with another individual about what constitutes the “correct” interpretation of that image.