A Novel Hybrid Approach for a Content-Based Image Retrieval Using Feature Fusion

Abstract

The multimedia content generated by devices and image processing techniques requires high computation costs to retrieve images similar to the user’s query from the database. An annotation-based traditional system of image retrieval is not coherent because pixel-wise matching of images brings significant variations in terms of pattern, storage, and angle. The Content-Based Image Retrieval (CBIR) method is more commonly used in these cases. CBIR efficiently quantifies the likeness between the database images and the query image. CBIR collects images identical to the query image from a huge database and extracts more useful features from the image provided as a query image. Then, it relates and matches these features with the database images’ features and retakes them with similar features. In this study, we introduce a novel hybrid deep learning and machine learning-based CBIR system that uses a transfer learning technique and is implemented using two pre-trained deep learning models, ResNet50 and VGG16, and one machine learning model, KNN. We use the transfer learning technique to obtain the features from the images by using these two deep learning (DL) models. The image similarity is calculated using the machine learning (ML) model KNN and Euclidean distance. We build a web interface to show the result of similar images, and the Precision is used as the performance measure of the model that achieved 100%. Our proposed system outperforms other CBIR systems and can be used in many applications that need CBIR, such as digital libraries, historical research, fingerprint identification, and crime prevention.

LJCD-Net: Cross-Domain Jamming Generalization Diagnostic Network Based on Deep Adversarial Transfer

“…The specific configurations of these comparison techniques are detailed below. M1: Joint fractional order Fourier transform and trap for jamming detection (FR-TRAP) scheme [ 6 ]; M2: Classical decision tree (DT) scheme based on thresholding [ 30 ]; M3: Time-domain constructed residual network suppression based jamming identification (TD-RNN) scheme [ 10 ]; M4: Deep subdomain adaptive network (DSAN) [ 11 ]; M5: DANN [ 31 ]. …”

“…The method converts the time domain signals into one-dimensional data as inputs to the network, but the amount of data is small and the network generalization ability is poor. However, the assumption of independent and identically distributed signals may frequently be violated in industrial application scenarios [ 11 , 12 ]. This is exemplified by the irregular relative motion between the interfering source and the localization receiver, which alters both the frequency and magnitude of jamming features.…”

A Novel Statistical Approach to Obtain the Best Visibility Slice in MRI Sequence of Brain Tumors

“…The authors Saha et al Clustering) for the purpose of analyzing tumor heterogeneity based on images 5 . The whole of the tumor voxel space is used to get the comprehensive tumor heterogeneity density profiles (THDPs), which are then analyzed 6 . Sikandar and Mahum 6 proposed a technique to address the problem of oversampling in MRI/CT scan image pre-processing.…”

An Investigation of Distance Measures for Development of Effective Content Based Tumor Image Retrieval System

“…Sikandar et al (10) devised a novel retrieval system that employs a transfer learning technique and incorporates ResNet50 and VGG16 pre-trained deep learning models, as well as one machine learning model, KNN. The test results were obtained with 100% precision.…”