Cbir Content Based Image Retrieval

3028 Words13 Pages

Abstract— CBIR is a technique that is growing interest in every field. CBIR is Content Based Image Retrieval which retrieves the image based on their content rather than keywords. This paper presents review of CBIR and its techniques. It gives an overview of currently available literature on CBIR. Different Distance Metrics like Euclidean distance, Manhattan Distance and Different feature extraction techniques like DCT, LBP are discussed in this paper
Index Terms— CBIR, Manhattan Distance, Euclidean Distance, LBP, DCT, SVM.
I. INTRODUCTION
CBIR stands for Content Based Image Retrieval. It is the process of searching image from large database using its content. Content can be color, shape, texture or any other spatial features. It is growing …show more content…

APPLICATIONS
CBIR is growing interest in many fields. Application areas of CBIR are as follows:

A. Crime Prevention
Records of criminals like photograph, fingerprint with other information is maintained by the police and investigating teams. These records are then used for face recognition and fingerprint matching.
To identify any person or to find if person is having the criminal record in past or not, these records are very useful and helpful. Using CBIR, new records are matched with previous records to find the information related to criminals.
In UK, this method is commonly in practice. Even Metropolitan Police Force in London involved with a project which is setting up an international database of images of stolen objects.

B. Medical Diagnosis
In medical field, visual information such as X- Rays, ultrasound, and scanned images with other details of patient is stored in the database. These records or images are helpful in finding the treatment with help of previous cases.
These records are also useful in teaching and research. With the help of these records, researchers can even find the treatment of new diseases. That’s why CBIR is growing interest in medical field also.

C. …show more content…

In this method, color histogram of each image in the database is computed. At the time of searching, color histogram of query image is also computed. Color histogram of query image is match with the color histogram of images in the database. Then those images are selected whose color histogram matches most closely with the histogram of query image. In this way, images are retrieved by calculating color histogram.

Fig. 5: Color Histogram of image

Color features method is widely used method for image retrieval. This method is easier as compared to shape and texture because color information is easy to extract as compared to texture and shape information. Extracting color features can be completed without regard to image size or orientation.
Color difference histogram is also used to retrieve the images from the database.

B. Shape
Shape does not refer to shape of image but it refers to shape of particular region. Shape can be determined by applying segmentation or edge detection.
Many methods use shape filters to identify given shapes of images. Some shape descriptors are Fourier transform and Moment

Open Document