Deep hash for latent image retrieval

Content based image retrieval systems (CBIR) are used to search images on the basis of their visual content in a huge image database. This approach uses a multi-clustering technique with a multi-searching process. In addition it integrate, Mean SIFT (Scale Invariant Feature Transform) descriptor as local feature and HSV(hue, saturation, value) histogram as global feature. Our proposition aims to a maximum separation of the execution to gain time and keep the performance of each descriptor. Local and global features are combined for more relevant results. Getting several views to relevant results to cover the subjectivity in displaying results. This article, detailed our proposed method with a comparison with the FIRE (Flexible Image Retrieval) Engine and LIRE (Lucene Image Retrieval). The results demonstrate the feasibility and relevance of our proposition.

The objective for this paper is to develop an adaptive image retrieval system with an innovative approach to use artificial neural networks and clustering techniques to retrieve images similar to the input image. This paper involves retrieving images from huge image databases which are visually similar to a query image. Due to the enormous increase in image database size, and its high usage in a variety of applications, need for the development of Content Based Image Retrieval arose. It summarizes the problem, the proposed solution, and the desired results. The system uses neural networks, relevance feedback and clustering. The system is made intelligent by making the system learn the user's preference as feedback. The relevance feedback is used to improve the precision of the system by analyzing user's relevance feedback for each retrieved image while neural network and clustering techniques are used to reduce the time complexity of the system. The system uses three-layered neural network to train the system using image clusters as training dataset by a supervised approach. Also after taking the feedback from the user, the image clusters are re-clustered by rearranging the images after a successful retrieval. Given a user query as an image, the neural system retrieves related images by computing similarities with images in the given image clusters. To provide preference, from all the retrieved images user selects an image as relevant one and all other are hence treated as irrelevant ones. So, the rank of the selected image is increased while the ranks of other images are decreased. With this feedback, the system refinement method estimates global approximations and adjusts the similarity probabilities.

Medical image indexing and storage is gaining increased importance. Indexing and retrieval of these images efficiently is becomes an essential task. Indexing of medical images using text or numbers is a cumbersome task, difficult to memorize and time consuming. Indexing and Retrieval of images can be done through query by text and query by image which is also known as content based image retrieval (CBIR). Different medical images possess different texture and shape features. This paper presents a novel and hybrid approach of managing a huge medical image database and retrieving a medical image from the database by fusion of shape and texture features. Image is retrieved by comparing the features of query image and images of database. We attach some document information with each medical image such as patient identity, diseases, age, and case history. It can be used for further diagnostic and analysis purpose. Further the proposed algorithm also updates the database automatically if new query is found. MATLAB ® 7.01 and its image processing toolbox have been used to implement the algorithm.

In this paper, a prototype system is described for the management and content-based retrieval of defect images in huge image databases. This is a real problem in surface inspection applications, since modern inspection systems may produce up to thousands of defect images in a day. The retrieval is based on shape and internal structure characteristics of defects, so neither manual labeling nor annotation needs to be made. We are using a noncommercial, generic content-based image retrieval (CBIR) system called PicSOM that is modified to fit the special requirements of our application. The results of experiments show that the prototype system works fast with good retrieval results.

Content Based Image Retrieval (CBIR) plays a significant role in identifying the similarity of images with large datasets. It is identified based on the size, colour, and texture features of the image. But in such conditions, it is complex to determine the features of query images in large datasets and does not show accurate similarity when compared with every image in the retrieval process. In order to perform an efficient similarity of images, a novel Machine Learning (ML) approach Kernelized Radial Basis Auto-Encoder Function Neural Network (Ker_RadBAEFNN) technique is proposed that performs the individual image classification in the retrieval process. Moreover, the neural networks are optimised based on the reinforcement process and perform the extraction process regarding individual images. Further,reinforcement-based optimisation estimates the images in neural networks for undertaking an automatic feature extraction of query images. The performance of the classification process is validated based on MNIST, METU, and COCO datasets that determined the efficiency of the recognition and classification process of image retrieval. The experimental analysis is carried out based on various measures such as accuracy, precision, recall, F1-score, RMSE, and MAPE for the proposed and existing GLCM-ABC, PSO-ANN, IRB-CNN, FAGWO, and OCAM methods. The analysis shows that the performance of the proposed attained better effectiveness with attained accuracy by 98% and diminished for state-of-the-art techniques as 92%, 95%, 94%, 96.8%, as well as 96%, respectively. Compared to existing methods, the accuracy rate of the proposed method is maximised by 1.3%.

A novel octopus based Parkinson’s disease and gender recognition method using vowels

Compressed domain image classification performs classification directly on compressive measurements acquired from the single-pixel camera, bypassing the image reconstruction step. It is of great importance for extending high-speed object detection and classification beyond the visible spectrum in a cost-effective manner especially for resource-limited platforms. Previous neural network methods require training a dedicated neural network for each different measurement rate (MR), which is costly in computation and storage. In this work, we develop an efficient training scheme that provides a neural network with dynamic-rate property, where a single neural network is capable of classifying over any MR within the range of interest with a given sensing matrix. This training scheme uses only a few selected MRs for training and the trained neural network is valid over the full range of MRs of interest. We demonstrate the performance of the dynamic-rate neural network on datasets of MNIST, CIFAR-10, Fashion-MNIST, COIL-100, and show that it generates approximately equal performance at each MR as that of a single-rate neural network valid only for one MR. Robustness to noise of the dynamic-rate model is also demonstrated. The dynamic-rate training scheme can be regarded as a general approach compatible with different types of sensing matrices, various neural network architectures, and is a valuable step towards wider adoption of compressive inference techniques and other compressive sensing related tasks via neural networks.

PurposeCurrent practices in data classification and retrieval have experienced a surge in the use of multimedia content. Identification of desired information from the huge image databases has been facing increased complexities for designing an efficient feature extraction process. Conventional approaches of image classification with text-based image annotation have faced assorted limitations due to erroneous interpretation of vocabulary and huge time consumption involved due to manual annotation. Content-based image recognition has emerged as an alternative to combat the aforesaid limitations. However, exploring rich feature content in an image with a single technique has lesser probability of extract meaningful signatures compared to multi-technique feature extraction. Therefore, the purpose of this paper is to explore the possibilities of enhanced content-based image recognition by fusion of classification decision obtained using diverse feature extraction techniques.Design/methodology/approachThree novel techniques of feature extraction have been introduced in this paper and have been tested with four different classifiers individually. The four classifiers used for performance testing were K nearest neighbor (KNN) classifier, RIDOR classifier, artificial neural network classifier and support vector machine classifier. Thereafter, classification decisions obtained using KNN classifier for different feature extraction techniques have been integrated by Z-score normalization and feature scaling to create fusion-based framework of image recognition. It has been followed by the introduction of a fusion-based retrieval model to validate the retrieval performance with classified query. Earlier works on content-based image identification have adopted fusion-based approach. However, to the best of the authors’ knowledge, fusion-based query classification has been addressed for the first time as a precursor of retrieval in this work.FindingsThe proposed fusion techniques have successfully outclassed the state-of-the-art techniques in classification and retrieval performances. Four public data sets, namely, Wang data set, Oliva and Torralba (OT-scene) data set, Corel data set and Caltech data set comprising of 22,615 images on the whole are used for the evaluation purpose.Originality/valueTo the best of the authors’ knowledge, fusion-based query classification has been addressed for the first time as a precursor of retrieval in this work. The novel idea of exploring rich image features by fusion of multiple feature extraction techniques has also encouraged further research on dimensionality reduction of feature vectors for enhanced classification results.

In this study, we will explore Neural Network based FPGA acceleration based on accelerating General Matrix Multiplication (GEMM). GEMM acceleration allows regularized and modular implementation of accelerator design, as well as providing the benefits of scalability. GEMM based designs also offer a degree of functional flexibility which is a key benefit to understand the highly dynamic architectural developments in Deep Learning algorithms. We quantify the theoretical performance model and tradeoffs of a GEMM accelerator along with exploration of the design space. Moreover, we propose a design for an accelerator exploiting 8-bit quantization to increase bandwidth while preserving model accuracy, exploiting FPGAs for model parallelization and data re-use for high performance and low latency neural network inference. Lastly, we test and evaluate our design on the MNIST dataset. The proposed method is useful to optimize the hardware area in Deep Learning systems without sacrificing performance.

The recognition of fully multi-oriented handwritten characters is a challenging problem. Contrary to univariate signals where the shift invariance property in the Fourier transform can be used, multivariate signals like images require special care to extract rotation invariant features. Several strategies to solve such classification tasks are possible. The proposed method considers input features obtained by the Radon transform or Polar transform. A convolutional neural network is then used for extracting higher level features. This classifier includes in addition the Fast Fourier Transform for extracting shift invariant features at the neural network level. The Radon transform and the convolutional layers process the image at the pixel level while the Fourier transform and the upper layers of the neural network process rotation invariant features. The classifier is evaluated on multi-oriented handwritten digits based on the MNIST database (Arabic digits) and on the ISI database (Bangla digits). The average recognition rate for multi-oriented characters is 93.10% for the Arabic digits and 77.01% for the Bangla digits. This neural architecture highlights the interest of the radial convolutional layer for the recognition of multi-oriented shapes.

In this chapter, the authors begin with what appears to be a short review of logistic regression but actually forms the basis of neural networks. They describe practical aspects of the implementation of neural networks in the traditional log-likelihood form and the new rank-based approach. Neural networks inherently have a large number of variables, which are collectively referred to as the feature space. This implies that a large number of parameters must be estimated using an optimization method. In order to optimize multi-layered neural networks, a gradient descent approach is used. The authors describe a well-known application of supervised learning for binary classification using neural networks, that of recognizing cats vs. dogs. They take on the more challenging problem of multi-class classification. The authors compare new rank-based neural network with log-likelihood neural network using the MNIST data set.

In recent years, secure image archival and retrieval model in cloud computing (CC) received significant attention to ensure data confidentiality and secured data transmission among cloud server and clients cloud storage and users. Traditional secure image retrieval (IR) techniques are not applicable to adapt with a large-scale IR in cloud environment. In order to overcome these issues, this article introduces an effective secure IR scheme using Inception with ResNet v2 (SIRS-IR) and Multiple Share Creation (MSC) scheme. The proposed method uses Inception with ResNet v2 model based feature extraction process. At the same time, the MSC process is utilized for multiple share generation and then encryption of shares takes place using double chaotic logistic map (DLCM) technique. Besides, the encrypted shares and the respective feature vectors are saved in the cloud server with the corresponding image identification number. During the IR process for the applied QI, the SIRS-IR model extracts the feature vectors and performs similarity measurement to retrieve the related images from the database interms of encrypted shares. Finally, the share decryption process is carried out for the reconstruction of original images with no loss of quality. Extensive experimentations were performed to verify the retrieval performance and image quality of the reconstructed images using Corel10K dataset. The obtained results stated that the presented SIRS-IR model is found to be superior to other methods in a considerable way.

In this paper we propose a fast and efficient image retrieval method for searching JPEG2000 compressed image databases. Comparing image contents of the JPEG2000 coded images in the pixel domain requires image decompression, which imposes intensive computational processes of the inverse discrete wavelet transform and the arithmetic decoding. On the other hand, the first decoding stage of the JPEG2000 standard is the packet header decoding, which is a simple process, but it provides valuable information about the code blocks in the packet. In this paper we exploit packet header information including the number of non-zero bit-planes, the number of coding passes and the code block length for comparing the JPEG2000 compressed images. Experimental results show that the proposed method provides a better performance compared to other JPEG2000 compressed domain retrieval techniques and even outperforms the pixel-based image retrieval techniques such as the Gabor filter; moreover the proposed method is very fast with very low computational load.

Deep hashing methods utilize an end-to-end framework to mutually learn feature representations and hash codes, thereby achieving a better retrieval performance. Traditional supervised hashing methods adopt handcrafted features for hashing function learning and then generate hash codes through classification and quantization. The lack of adaptability and independence of the quantization procedure leads to low retrieval accuracy of supervised hashing methods with handcrafted features in image retrieval. In this study, a non-relaxation deep hashing method for fast image retrieval is proposed. In this method, a differentiable host thresholding function is used to encourage hash-like codes to approach -1 or 1 non-linearly at the output of the convolutional neural, instead of the symbol function for quantization used in the traditional method. The output of the host thresholding function is directly used to compute the network training error, and a loss function is elaborately designed with the norm to constrain each bit of the hash-like code to be as binary as possible. Finally, a symbol function is added outside the trained network model to generate binary hash codes for image storage and retrieval in a low-dimensional binary space. Extensive experiments on two large-scale public datasets show that our method can effectively learn image features, generate accurate binary hash codes, and outperform state-of-the-art methods in terms of the mean average precision.

Hierarchical Autoassociative Polynimial Network (HAP Net) for pattern recognition