70 research outputs found
Color Image Compression Using Vector Quantization and Hybrid Wavelet Transform
No full textAbstractThis paper presents simpler image compression technique using vector quantization and hybrid wavelet transform. Hybrid wavelet transform is generated using Kronecker product of two different transforms. Image is converted to transform domain using hybrid wavelet transform and very few low frequency coefficients are retained to achieve good compression. Vector quantization is applied on these coefficients to increase compression ratio significantly. VQ algorithms are applied on transformed image and codebooks of minimum possible size 16 and 32 are generated. KFCG and KMCG are faster in execution and beats performance of LBG algorithm. KFCG combined with hybrid wavelet transform gives lowest distortion and acceptable image quality at compression ratio 192
Modified CSLBP
No full textImage hashing is an efficient way to handle digital data authentication problem. Image hashing represents quality summarization of image features in compact manner. In this paper, the modified center symmetric local binary pattern (CSLBP) image hashing algorithm is proposed. Unlike CSLBP 16 bin histogram, Modified CSLBP generates 8 bin histogram without compromise on quality to generate compact hash. It has been found that, uniform quantization on a histogram with more bin results in more precision loss. To overcome quantization loss, modified CSLBP generates the two histogram of a four bin. Uniform quantization on a 4 bin histogram results in less precision loss than a 16 bin histogram. The first generated histogram represents the nearest neighbours and second one is for the diagonal neighbours. To enhance quality in terms of discrimination power, different weight factor are used during histogram generation. For the nearest and the diagonal neighbours, two local weight factors are used. One is the Standard Deviation (SD) and other is the Laplacian of Gaussian (LoG). Standard deviation represents a spread of data which captures local variation from mean. LoG is a second order derivative edge detection operator which detects edges well in presence of noise. The proposed algorithm is resilient to the various kinds of attacks. The proposed method is tested on database having malicious and non-malicious images using benchmark like NHD and ROC which confirms theoretical analysis. The experimental results shows good performance of the proposed method for various attacks despite the short hash length
Architecture of a programmable system-on-chip platform for flexible radio processing
No full textThe emergence of multiple radio access technologies (RATs) and their continuous evolution, is driving the need for programmable radio processing. Programmable radio devices with run-time flexibility and resource virtualization features will not only enable faster time-to-market, longer lifetime of devices, and universal connectivity, but also act as building blocks for advanced wireless technologies of adaptive and cognitive radios. These requirements have forced a shift from the traditional ASIC approach. However, most existing flexible solutions are based on either fully software-defined or software-controlled approaches that lack the power efficiency, performance and determinism (for real-time constraints) needed for wireless processing. In this thesis, we propose a programmable multi-processor system-on-chip (SoC) platform architecture based on a novel Virtual Flow Pipelining (VFP) framework that aims at striking a balance between flexibility (as provided by SDR) and performance (as provided by ASICs). The key highlights of this concept are a simple task-level programming model for provisioning protocol flows, and the use of dedicated hardware-based OS-like support for controlling their run-time execution. We present the evolution of a clustering-based organization for the SoC with distributed-shared controllers. Clustering along with an inherent architectural support for message passing provides a balance between scalability and hardware overhead. Shared controllers with a pipelined microarchitecture and a separate interconnect for control messaging are designed for low hardware complexity and high performance. The proposed architecture is evaluated by creating a bit- and cycle-accurate model in synthesizable register-transfer-level (RTL). It has been built into a virtual platform for 802.11a transmitter, which has successfully executed single and multiple flows for rates of 6, 12 and 24 Mbps. This thesis also presents a characterization and analysis of the architecture to provide key implications such as control overhead for different task sizes, its impact on cluster size etc.M.S.Includes bibliographical referencesby Onkar Sarod
Adaptive CSLBP compressed image hashing
No full textHashing is popular technique of image authentication to identify malicious attacks and it also allows appearance changes in an image in controlled way. Image hashing is quality summarization of images. Quality summarization implies extraction and representation of powerful low level features in compact form. Proposed adaptive CSLBP compressed hashing method uses modified CSLBP (Center Symmetric Local Binary Pattern) as a basic method for texture extraction and color weight factor derived from L*a*b* color space. Image hash is generated from image texture. Color weight factors are used adaptively in average and difference forms to enhance discrimination capability of hash. For smooth region, averaging of colours used while for non-smooth region, color differencing is used. Adaptive CSLBP histogram is a compressed form of CSLBP and its quality is improved by adaptive color weight factor. Experimental results are demonstrated with two benchmarks, normalized hamming distance and ROC characteristics. Proposed method successfully differentiate between content change and content persevering modifications for color images
Abfraction Lesions: Analyzing the cause by identifying biological and lifestyle similarities
No full textAbfraction lesions are non-carious (not caused by decay), wedge-shaped indentations near the gumline that expose the sensitive dentin under the protective enamel (Sarode and Sarode, 2013). A clear understanding of the contributors to abfraction formation would have a major impact on clinical dentistry, potentially informing interventions to prevent these lesions. Knowing the causes of abfractions would also benefit forensic anthropology and bioarchaeology, since dental structures are widely used to determine aspects of a deceased individual’s identity, including their age, sex, population affinity, habits, and general lifestyle. However, there little consensus on the etiology of abfractions among the various disciplines affected (Jakupovic et al., 2014; Nascimento et al., 2016)
A framework for cloud cover prediction using machine learning with data imputation
No full textThe climatic conditions of a region are affected by multiple factors. These factors are dew point temperature, humidity, wind speed, and wind direction. These factors are closely related to each other. In this paper, the correlation between these factors is studied and an approach has been proposed for data imputation. The idea is to utilize all these features to obtain the prediction of the total cloud cover of a region instead of removing the missing values. Total cloud cover prediction is significant because it affects the agriculture, aviation, and energy sectors. Based on the imputed data which is obtained as the output of the proposed method, a machine learning-based model is proposed. The foundation of this proposed model is the bi-directional approach of the long short-term memory (LSTM) model. It is trained for 8 stations for two different approaches. In the first approach, 80% of the entire data is considered for training and 20% of the data is considered for testing. In the second approach, 90% of the entire data is accounted for training and 10% of the data is accounted for testing. It is observed that in the first approach, the model gives less error for prediction
- …
