Medical image retrieval using deep convolutional neural network

In this paper, a new image indexing and retrieval algorithm using local mesh patterns are proposed for biomedical image retrieval application. The standard local binary pattern encodes the relationship between the referenced pixel and its surrounding neighbors, whereas the proposed method encodes the relationship among the surrounding neighbors for a given referenced pixel in an image. The possible relationships among the surrounding neighbors are depending on the number of neighbors, P. In addition, the effectiveness of our algorithm is confirmed by combining it with the Gabor transform. To prove the effectiveness of our algorithm, three experiments have been carried out on three different biomedical image databases. Out of which two are meant for computer tomography (CT) and one for magnetic resonance (MR) image retrieval. It is further mentioned that the database considered for three experiments are OASIS-MRI database, NEMA-CT database, and VIA/I-ELCAP database which includes region of interest CT images. The results after being investigated show a significant improvement in terms of their evaluation measures as compared to LBP, LBP with Gabor transform, and other spatial and transform domain methods.

Local Mesh Patterns Versus Local Binary Patterns: Biomedical Image Indexing and Retrieval

Local ternary co-occurrence patterns: A new feature descriptor for MRI and CT image retrieval

Convolutional neural network (CNN) gained great attention for robust feature extraction and information mining. CNN had been used for variety of applications such as object recognition, image super-resolution, semantic segmentation etc. due to its robust feature extraction and learning mechanism. By keeping constant the baseline learning topology, various CNN architectures were proposed to improve the respective system performance. Among these, AlexNet, VGG16 and VGG19 are the famous CNN architecture introduced for object recognition task. In this paper, we make use of transfer learning to fine-tune the pre-trained network (VGG19) parameters for image classification task. Further, performance of the VGG 19 architecture is compared with AlexNet and VGG16. Along with the CNN architectures, we have compared the hybrid learning approach which is comprised of robust feature extraction from CNN architecture followed by support vector machine (SVM) classifier. We have used two state-of-the-art databases namely: GHIM10K and CalTech256 to study the effect of CNN architecture for robust feature extraction. Performance evaluation has been carried out using average recall, precision and F-score. Performance analysis shows that fine-tuned VGG19 architecture outperforms the other CNN and hybrid learning approach for image classification task.

Transfer Learning for Image Classification

Fusion of color histogram and LBP-based features for texture image retrieval and classification

Local maximum edge binary patterns: A new descriptor for image retrieval and object tracking

A new image indexing and retrieval system for content based image retrieval (CBIR) is proposed in this paper. The characteristics (vector points) of image are computed using color (color histogram) and SOT (spatial orientation tree). The SOT defines the spatial parent-child relationship among wavelet coefficients in multi-resolution wavelet sub-bands. First the image is divided into sub-blocks and then constructed the SOT for each low pass wavelet coefficient is considered as a vector point of that particular image. Similarly the color histogram features are collected from the each sub-block. The vector points of each image are indexed using vocabulary tree. The retrieval results of the proposed method are tested on different image databases, i.e., natural image database consists of Corel 1000 (DB1), Brodatz texture image database (DB2) and MIT VisTex database (DB3). The results after being investigated show a significant improvement in terms of average precision, average recall and average retrieval rate on DB1 database and average retrieval rate on texture databases (DB2 and DB3) as compared with most of existing techniques on respective databases.

Expert system design using wavelet and color vocabulary trees for image retrieval

In this paper, a new image-indexing algorithm with the combination of wavelet and Rotated Wavelet Correlogram (RWC) is proposed in contrast to the Gabor Wavelet Correlogram (GWC) The 0° and 90° information (sub-band) of image collected from Standard Wavelet Filters (SWFs) and +45° and −45° are collected from Rotated Wavelet Filters (RWFs) and further used for correlogram feature calculation The data handling with proposed method is very less compared with GWC therefore this method is six times faster than the GWC, which make it suitable for online application The retrieval results of the proposed method are tested on three different databases, ie, Corel 1000 (DB1) natural image database, MIT VisTex (DB2) texture image database and Corel 2450 (DB3) natural image database The results after being investigated show a significant improvement in terms of their evaluation measures as compared with most of existing techniques on respective databases

M. Subrahmanyam

Papers

Local maximum edge binary patterns: A new descriptor for image retrieval and object tracking

Expert system design using wavelet and color vocabulary trees for image retrieval

A correlogram algorithm for image indexing and retrieval using wavelet and Rotated Wavelet Filters