Online Store - 8925533488 /89

Chennai - 8925533480 /81

Hyderabad - 8925533482 /83

Vijayawada -8925533484 /85

Covai - 8925533486 /87

Text Based Image Retrieval Using Neural Network Algorithm

( 0 Rating )
Shape Image One
0 student


We gift a image retrieval system supported a combined search of text and content. But it has been going complex to retrieve the image from text. We have so many implementations regarding the content based image retrieval. In this project the concept is to use the text gift in title, description, and tags of the photographs for improving the results obtained with a customary text-based search. the system retrieves similar pictures from the gathering, and a complicated search designed for skilled users, wherever the space functions and weights are often bespoken. Here we are using one of the most important algorithm t train the data, that is Neural Network algorithm. Finally we get the data(images) based on the text which we given.



In recent years, collections of digital pictures are created and increased  chop-chop. In several areas of academe, commerce, government, medicine,  and net, a large quantity of data is out there. However, we have a tendency to cannot access or build use of this data unless it is organized to permit efficient browsing, searching, and retrieval. during all the most issues is that the problem of locating a desired image in a giant and varied assortment. whereas it is perfectly possible to identify a desired image from a tiny low assortment just by browsing, simpler techniques are required with collections containing thousands of things. Image retrieval attracts interest among researchers within the fields of image process, multimedia, digital libraries, remote sensing, astronomy, database applications, and different connected areas.  Image retrieval has been a really active analysis space since Nineteen Seventies, with the thrust of 2 major analysis communities: management and pc vision.  Therefore, image retrieval is outlined as  the task of finding out pictures  in a picture information.  image retrieval techniques is classified into 3 categories: text-based image retrieval, content-based image retrieval and semantic-based image retrieval. consecutive subsections can discuss these classes in additional detail. Text-based search provides results with linguistics similarity, while content-based search provides results with visual similarity. because of the independence between these approaches, is probably going that their combination may improve the performance of a groundwork system by benefiting of each approaches. within the gift work, we tend to gift a image retrieval system supported search of text.


Survey of text based image retrieval:

TBIR are often copied back to the late Nineteen Seventies. a really fashionable framework of TBIR was initial annotated the photographs by text then used text-based management systems to perform image retrieval. TBIR is accustomed manually annotate the image within the information with annotations, keywords, or descriptions.This method is accustomed describe each image contents and alternative data of the image such as: image file name, image and  image format, image size,and  image dimensions. Then, the user formulates matter or numeric queries to retrieve all pictures that area unit satisfying a number of the


International Journal of laptop and knowledge Technology (ISSN: 2279 – 0764)  Volume 04 – Issue 01, January 2015 fifty nine  criteria supported these annotations, However, there area unit some drawbacks in TBIR . the primary downside is that the foremost descriptive annotations should sometimes be entered manually. Manually annotation for an out sized image information is impractical. The second downside is that the foremost pictures area unit terribly made in its content and has a lot of details. The commentator could provide completely different descriptions to photographs with similar visual contents. Also, matter annotations area unit language-dependent.


Existing System:

Existing means previous works which we already done. By referencing some papers we got the information regarding content based image retrieval(CBIR) and semantic based image retrieval(SBIR).



  • Accuracy is not up to maximum or accuracy.
  • Continuous filtering is needed for achieving accurate results.
  • Data base is limited to acquire the image from large amount of data.


Proposed  Methodology:

 Block  Diagram:

Image Retrieval Using Neural Network Algorithm


1 Input Image Data set

2 Local Binary Pattern(LBP)

3 Discrete Wavelet Transform(DWT)

4 GLCM feature extraction

5 NN(Neural Network) algorithm


Work flow:


Conclusion and Future Scope:

In this project we have done successfully implemented the text based image retrieval system based on the neural network training system. To calculate the features of the data set and to retrieve the images we are used LBP and DWT techniques.

Another technique for image retrieval used to integrate text and image content to enhance the retrieval accuracy. Both the text and content-based techniques have their own characteristics, advantages, and disadvantages. By combining them, parts of their disadvantages can be overcome.



[1] Yakhnenko and V. Honavar, “Annotating images and image objects  using a hierarchical dirichlet process model,” Proceedings of the 9th  International Workshop on Multimedia Data Mining: held in  conjunction with the ACM SIGKDD 2008, ACM, 2008, pp. 1–7.

[2] Klimis S. Ntalianis Dionyssios D. Sourlas Konstantinos A. Raftopoulos  and S.D. Kollias, “Mining User Queries with Markov Chains,”  Application to Online Image Retrieval IEEE Transactions On  Knowledge And Data Engineering, vol. 25, No. 2, 2013.

[3] N.V. A.B. Chan P.J. Moreno G. Carneiro, “Supervised Learning of  Semantic Classes for Image Annotation and Retrieval,” Pattern Analysis  and Machine Intelligence, IEEE Transactions, vol. Vol.29,No.3, 2007,  pp. 394–410.

[4] S.H. Yang, J. Bian, and H. Zha, “Hybrid generative/discriminative  learning for automatic image annotation,” arXiv preprint  arXiv:1203.3530, 2012.

[5] E. Nowak, F. Jurie, and B. Triggs, “Sampling strategies for bag-of- features image classification,” Computer Vision–ECCV 2006, Springer,  2006, pp. 490–503.

[6] J. Verbeek, M. Guillaumin, T. Mensink, and C. Schmid, “Image  annotation with tagprop on the MIRFLICKR set,” Proceedings of the  international conference on Multimedia information retrieval, ACM,  2010, pp. 537–546.

[7] J.-H. Su, C.-L. Chou, C.-Y. Lin, and V.S. Tseng, “Effective image  semantic annotation by discovering visual-concept associations from  image-concept distribution model,” Multimedia and Expo (ICME), 2010  IEEE International Conference on, IEEE, 2010, pp. 42–47.

[8] A. Makadia, V. Pavlovic, and S. Kumar, “Baselines for image  annotation,” International Journal of Computer Vision, vol. 90, 2010, pp.  88–105.

[9] M. Everingham, L. Van Gool, C.K. Williams, J. Winn, and A.  Zisserman, “The pascal visual object classes (voc) challenge,”  International journal of computer vision, vol. 88, 2010, pp. 303–338.

[10] H. Bay, T. Tuytelaars, and L. Van Gool, “Surf: Speeded up robust  features,” Computer Vision–ECCV 2006, Springer, 2006, pp. 404–417.

[11] E. Rublee, V. Rabaud, K. Konolige, and G. Bradski, “ORB: an efficient alternative to SIFT or SURF,” Computer Vision (ICCV), 2011 IEEE  International Conference on, IEEE, 2011, pp. 2564–2571.

[12] G. ThakoreDarshak “Evaluation enhancement development and  implementation of content based image retrieval  algorithms”.PhDThesis.Maharaja Sayajirao University .2013.

Curriculum is empty

pantech team

Agile Project Expert

Course Rating

0.00 average based on 0 ratings

Course Preview
  • Price
  • Instructor pantech team
  • Duration 15 Hrs
  • Enrolled 0 student
  • Access 3 Months

More Things You Might Like This


Student Performance Prediction using Machine Learning

Abstract: Although the educational level of the Portuguese population has improved in the last decades, the statistics keep Portugal at Europe’s tail end due to its high student failure rates. In particular, lack of success in the core classes of Mathematics and the Portuguese language is extremely serious. On the other hand, the fields of


Student feedback analysis

Abstract: Advances in natural language processing (NLP) and educational technology, as well as the availability of unprecedented amounts of educationally-relevant text and speech data, have led to an increasing interest in using NLP to address the needs of teachers and students. Educational applications differ in many ways, however, from the types of applications for which


Machine Learning based Regression Model for Prediction of Soil Surface Humidity over Moderately Vegetated Fields

Abstract: Agriculture is one of the major revenue producing sectors of India and a source of survival. Numerous seasonal, economic and biological patterns influence the crop production but unpredictable changes in these patterns lead to a great loss to farmers. These risks can be reduced when suitable approaches are employed on data related to soil

Open Whatsapp Chat
Need Any Help?
Welcome to Pantech eLearning!..

How can i help you?