Analysis of KNN and SVM Classifier for Image Classification

As the growth and development of various multimedia technologies in the field of CBIR many advanced information retrieval systems have become popular and has brought the new evolution in fast and effective retrieval. In this paper the techniques of image classification in CBIR are been discussed and compared. It also introduces classifiers like support vector machine, Bayesian classifier for accurate and efficient retrieval of images.

Content Based Image Retrieval (CBIR) techniques are becoming an essential requirement in the multimedia systems with the widespread use of internet, declining cost of storage devices and the exponential growth of un-annotated digital image information available in recent years. Therefore multi query systems have been used rather than a single query in order to bridge the semantic gaps and in order to understand user’s requirements. Moreover, query replacement algorithm has been used in the previous works in which user provides multiple images to the query image set referred as representative images. Feature vectors are extracted for each image in the representative image set and every image in the database. The centroid, Crep of the representative images is obtained by computing the mean of their feature vectors. Then every image in the representative image set is replaced with the same candidate image in the dataset one by one and new centroids are calculated for every replacement .The distance between each of the centroids resulting from the replacement and the representative image centroid Crep is calculated using Euclidean distance. The cumulative sum of these distances determines the similarity of the candidate image with the representative image set and is used for ranking the images. The smaller the distance, the similar will be the image with the representative image set. But it has some research gaps like it takes a lot of time to extract feature of each and every image from the database and compare our image with the database images and complexity as well as cost increases. So in our proposed work, the KNN algorithm is applied for classification of images in the database image set using the query images and the candidate images are reduced to images returned after classification mechanism which leads to decrease the execution time and reduce the number of iterations. Hence due to hybrid model of multi query and KNN, the effectiveness of image retrieval in CBIR system increases. The language used in this work is C /C++ with Open CV libraries and IDE is Visual studio 2015. The experimental results show that our method is more effective to improve the performance of the retrieval of images.

A Content Based Image Retrieval System (CBIR) is proposed using features representing the images quite uniquely. Two algorithms have been used to extract two types of features from each of the images stored in the database. They are Modified Block Truncation Coding (MBTC) and Keri's Fast Codebook Generation (KFCG) algorithm. The features obtained from the images using the two algorithms represent each of them with unique and different values. KFCG algorithm forms a codebook containing codebook vectors. Codebook vectors are a set of code words used to encode the images. KFCG algorithm requires lesser time for vector quantization process. MBTC features can only represent images and hence useful for retrieval of images from database. MBTC features consist of upper and lower mean values of three colour components of colour space of pixels in the colour images. These features yield high performance for the CBIR system proposed in this paper. The images are analysed, compared and retrieved from the database using one of the two types of features. The analysis and comparison is also done using two different methods. Euclidean Distance is used to identify the image being enquired using two different types of features. Similarly, Support Vector Machine (SVM) is also used to obtain the result using two different types of features. SVM is implemented in two stages. First, the features are used to train the SVM variables. During training, some of the image features of different categories being utilized for training purpose are marked by users according to their categories. In the next stage, the SVM variables obtained during training are used to classify the images. A database of 1000 images of 10 different categories are used to perform assessment of the system. The system is implemented using MATLAB.

Retrieval of an image is a more effective and efficient for managing extensive image database. Content Based Image Retrieval (CBIR) is a one of the image retrieval technique which uses user visual features of an image such as color, shape, and texture features etc. It permits the end user to give a query image in order to retrieve the stored images in database according to their similarity to the query image. In this work, content based image retrieval is accomplished by combining the two features such as color and texture. Color features are extracted by using hsv histogram, color correlogram and color moment values. Texture features are extracted by Segmentation based Fractal Texture Analysis (SFTA). The combined features which are made up of 32 histogram values,64 color correlogram values, 6 color moment values and 48 texture features are extracted to both query and database images. The extracted feature vector of the query image is compared with extracted feature vectors of the database images to obtain the similar images. The main objective this work is classification of image using SVM algorithm.

In this technology era, images have become a major part of information processing. An Image plays an important role in Image registration (IR) processing for the extraction of information. There are various fields like in medical, tourism and geological, weather systems forecasting used image registration. In this paper, IR is presented based on Support Vector Machine learning in the content-based image retrieval system. A Support Vector Machine (SVM) for the purpose of retrieval of images similar to the query image. Using the SVM classifier, the system can retrieve more images relevant to the query in the database efficiently. There are many traditional techniques that have been used to retrieve images. One of the Content-based image retrievals has the most popular research area in the last few years. Image retrieval is a technique of finding out the most important features of the image. The main task of content-based image retrieval (CBIR) is to get a similar images as well as perfect and fast result. In this CBIR system, effective organization of the image database used to improve the performance of the system. The study of contentbased image retrieval (CBIR) technique has become an important research issue. In this way, studied and analyzed of various features as an individual or in combinations. Through the studied of various research papers after that conclude the color and texture-based feature extraction is the most important for imparting the best extraction and support vector machine makes this task more easy and effective.

The dramatic rise in the sizes of images databases has stirred the development of effective and efficient retrieval systems. The development of these systems started with retrieving images using textual connotations but later introduced image retrieval based on content. This came to be known as Content Based Image Retrieval or CBIR. Systems using CBIR retrieve images based on visual features such as texture, color and shape, as opposed to depending on image descriptions or textual indexing. In the proposed work we will use various types of image features like color, texture, shape, energy, amplitude and cluster distance to classify the images according to the query image. We will use multi-SVM technique along with clustering technique to compare the features of the input image with the input dataset of images to extract the similar images as that of the query image

A simple and efficient CBIR system with good retrieval accuracy is designed without using any intensive image processing feature extraction techniques. The unique indexed color histogram and wavelet decomposition based horizontal, vertical and diagonal image attributes (WH) serve as the main features for the retrieval system. Support vector machine (SVM) and decision tree (DT) is used for classification of these distinct image features to further improve the retrieval accuracy of the system. The performance of the proposed content based image classification and retrieval systems are evaluated with the standard SIMPLIcity dataset. The performance of the system is measured with precision as the metric. K-fold cross validation is used for validating the results. The proposed system performs obviously better than SIMPLIcity and all the other compared methods like FEI, FIRM and Simple Histogram.

2012

As the growth and development of various multimedia technologies in the field of CBIR many advanced information retrieval systems have become popular and has brought the new evolution in fast and effective retrieval. In this paper the techniques of image classification in CBIR are been discussed and compared. It also introduces classifiers like support vector machine, Bayesian classifier for accurate and efficient retrieval of images. KeywordsContent Based Image Retrieval, RLS classifier, Bayesian classifier, support vector machine, relevance feedback.

INTERNATIONAL JOURNAL OF ADVANCE RESEARCH, IDEAS AND INNOVATIONS IN TECHNOLOGY

The dramatic rise in the sizes of pictures databases has blended the advancement of powerful and productive recovery frameworks. The improvement of these frameworks began with recovering pictures utilizing printed implications however later presented picture recovery dependent on the substance. This came to be known as Content-Based Image Retrieval or CBIR. Frameworks utilizing CBIR recover pictures dependent on visual highlights, for example, surface, shading and shape, rather than relying upon picture depictions or printed ordering. In the proposed work we will use various types of image features like colour, texture, shape, energy, amplitude and cluster distance to classify the images according to the query image. We will use multi-SVM technique along with the clustering technique to compare the features of the input image with the input dataset of images to extract the similar images as that of the query image.

Journal of Physics: Conference Series

Retrieving visually similar images from image database needs high speed and accuracy. Various text and content based image retrieval techniques are being investigated by the researchers in order to exactly match the image features. In this paper, a content-based image retrieval system (CBIR), which computes color similarity among images, is presented. CBIR is a set of techniques for retrieving semantically relevant images from an image database based on automatically derived image features. Color is one important visual features of an image. This document gives a brief description of a system developed for retrieving images similar to a query image from a large set of distinct images with histogram color feature based on image index. Result from the histogram color feature extraction, then using k-means clustering to produce the image index. Image index used to compare to the histogram color feature of query image and thus, the image database is sorted in decreasing order of similarity. The results obtained by the proposed system obviously confirm that partitioning of image objects helps in optimization retrieving of similar images from the database. The proposed CBIR method is compared with our previously existed methodologies and found better in the retrieval accuracy. The retrieval accuracy are comparatively good than previous works proposed in CBIR system.

Journal of advance research in dynamical and control system, 2019

Content based Remote Sensing Image Retrieval (CBRSIR) systems are used to retrieve useful contents from a massive amount of remote sensing images. A proposed novel technique for remote sensing image retrieval using SVM (Support Vector Machine) classifier with multi-features such as HOG (Histogram of Oriented Gradients), Color moment, Gabor and wavelet. Initially, the color feature is extracted from the satellite image using the peculiar color moments and are invariant to scaling and rotation. Texture features are extracted using dominant HOG, Gabor and Wavelet then the feature selection methods are separately classified. In the existing system, the images are first retrieved and then classified. Moreover the retrieval rates obtained by the existing techniques are not satisfactory. Hence a method is proposed as a novel multi feature based SVM technique where it initially classifies the different class from the data base and get the similar retrieval images based upon the feature of the query image. Thus by finding the retrieval rate after performing the classification, it is evident that output retrieval rate is better by comparing with other models. The fundamental performance metrics like accuracy, sensitivity and specificity are taken into comparison. The proposed method has higher accuracy when it is compared to the accuracy of other feature based SVM.

Due to the enormous increase in image database sizes, the need for an image search and indexing tool is crucial. Content-based image retrieval systems (CBIR) have become very popular for browsing, searching and retrieving images in different fields including web based searching, industry inspection, satellite images, medical diagnosis images, etc. The challenge, however, is in designing a system that returns a set of relevant images i.e. if the query image represents a horse then the first images returned from a large image dataset must return horse images as first responses. In this paper, we have combined YACBIR [7], a CBIR that relies on color, texture and points of interest and Multiple Support Vector Machines Ensemble to reduce the existing gap between high-level semantic and low-level descriptors and enhance the performance of retrieval by minimize the empirical classification error and maximize the geometric margin classifiers. The experimental results show that the method proposed reaches high recall and precision.

Traditionally, the image is retrieved using keywords, but to assign the keywords to large amount of database increase the man efforts. To avoid such situation, CBIR becomes the popular research area. In CBIR, the image is retrieved based on the content. The content of the image is determined by fetching its different features. Basically any image consists of different shapes, colors and textures. So these properties can be taken as the features of the image. One query image is given as a input and from that image, the similar images are extracted from database. In this paper, color features are extracted and compared to find the similar images from database. It is tested on different type of images. Precision and recall are used as comparison parameters to evaluate the results. Here, the precision rate is 87% and recall rate is 79%.

arXiv (Cornell University), 2013

Feature means countenance, remote sensing scene objects with similar characteristics, associated to interesting scene elements in the image formation process. They are classified into three types in image processing, that is low, middle and high. Low level features are color, texture and middle level feature is shape and high level feature is semantic gap of objects. An image retrieval system is a computer system for browsing, searching and retrieving images from a large image database. Content Based Image Retrieval (CBIR) is a technique which uses visual features of image such as color, shape, texture, etc…to search user required image from large image database according to user's requests in the form of a query. MKNN is an enhancing method of KNN. The proposed KNN classification is called MKNN. MKNN contains two parts for processing, they are validity of the train samples and applying weighted KNN. The validity of each point is computed according to its neighbors. In our proposal, Modified K-Nearest Neighbor (MKNN) can be considered a kind of weighted KNN so that the query label is approximated by weighting the neighbors of the query. The procedure computes the fraction of the same labeled neighbors to the total number of neighbors. MKNN classification is based on validated neighbors who have more information in comparison with simple class labels. This paper also concentrates identifying the unlabeled images with help of MKNN algorithm. Experiments show the validity takes into accounts the value of stability and robustness of the any train samples regarding with its neighbors and excellent improvement in the performance of KNN method. This system allows provide label to unlabeled image as user input.

In this paper, we propose content-based image retrieval method based on the SVM approach with efficient combination of histogram, color and edge features. Thus extending the work of the previous approach which used same set of features and the Euclidean distance measurement technique. As its histogram features, the extracted histogram bar values for each and every image are used. As its color features, the image is segmented into small pieces and then for each piece the red, green, blue values are used .As its edge features, Canny’s edge detection technique is used to extract the maximum edge value of the image. After extracting the features, we use a machine learning technique called SVM (support vector machine) to find out the optimal result. Combining the features and classifying them using SVM and then finally comparing the results with the previous approach not only gives a better accuracy, but also evaluates the generalization ability under the limited training samples. The analysis of the proposed work is done using MATLAB R2007b simulator.