CONTENT BASED IMAGE
RETRIEVAL USING FEATURE
EXTRACTION WITH
CONVOLUTIONAL NERUAL
NETWORKS
CONTENTS
ABSTRACT
INTRODUCTION
EXISTING METHODS
PROPOSED METHODS
COMPARISION
ADVANTAGES AND DISADVANTAGES
APPLICATIONS
RESULT
CONCLUSION
Abstract
The retrieval performance of a content-based image retrieval (CBIR) system is mainly influenced by
the feature representations and similarity measures. Recently, deep learning develops rapidly and the deep
features based on deep learning have been applied widely because it has been shown that the features
have very strong generalization. This paper applies the original deep features generated by convolution
neural network (CNN) to CBIR, which can separate similar image pairs and dissimilar image pairs to a
large degree. The input in this paper is pair features which are assembled by pair of images: the query
image and each test image in the image dataset. The test images then are ranked by the distance between
the pair features and the trained hyperplane. Experiments show that our method can significantly improve
the performance of CBIR for object image retrieval tasks.
INTRODUCTION
A very large amount of data is available and with each passing day gigabytes of
information is generated. A user is much interested to find their relevant information
such as images from those multimedia databases.
The user query image features are matched with the stored feature database that
gives us the best possible match according to the user query image.
Several similarity matching methods have been proposed which usually calculate
the distance between the query image and the database images.
The basic use of SIFT algorithm is that this method gives us the better results than
other feature extraction techniques.
This algorithm is invariant to rotation, scaling methods so we need not to worry
about these issues the only thing of concern is the illumination and camera point of
view.
EXISTING
SYSTEM
As stated earlier CBIR is a retrieval
technique of images from large
image database based on visual
features of an image such as colour,
shape and texture. The basic concept
used in this paper involves the color
feature, shape feature and also
texture feature for retrieval of
images.
Algorithm
Read Read the image dataset.
Extract Extract the features using SIFT method.
Optimize Optimize the feature set by using optimization algorithm (BFOA).
Classify Classify the optimized data using DNN.
Extract Extract the features of the query image (RGB).
Optimize Optimize the query image features.
Classify Classify features from the feature database.
Categorize Categorize the calculations based on both the images query image and database image.
PROPOSED SYSTEM
Content Based Image Retrieval using Feature Extraction
with Convolution Nerual Network
Results
RESULTS
Retrieving of 10 images from the content based efficient visual features
The content-based image retrieval is used for the retrieving 10 images where the retravel of data is mainly based on
the CNN classifier here the use of CBIR images in an best possible way to address the similarity in less amount of
time.
Advantages and Disadvantages
Conclusion
Finally have conducted other experiments such as Landmark Image Retrieval, but our
method performs best in many parameters. As far as we concerned, the huge gap between the
landmark datasets and Imagenet, which is the training dataset of the pre-trained CNN, leads to
the poor learning because it may not be able to fill the gap. Hence, we try to CNN to replace
the linearity in, but the retrieval processing turns out to be inefficient in previous methods
compared to proposed method. However, improving the generalization of our method remains
to be a challenge for us in future in large datasets.
Thank you
