Search for similar images with images (visual words)

Last Update:2015-03-06 Source: Internet

Author: User

Tags install matplotlib

Developer on Alibaba Coud: Build your first app with APIs, SDKs, and tutorials on the Alibaba Cloud. Read more ＞

Code address

Code GitHub Address

Preparing files

Vlfeat (open source implementation of SIFT):
The decompression will be vlfeat-0.9.20\bin\win32 added to the path path of the system in order to use the SIFT command at the command line
Install Pysqlite:pip install pysqlite
Install Matplotlib:pip isntall matplotlib
Install CherryPy (Lightweight Web server):pip install cherrypy
Photo Gallery (cat cat): 10000 cat pictures

Process

Extracting feature points for each picture with Vlfeat (Sift feature model)
To sample the feature points of each graph and cluster them into a visual word of a picture
That is, the visual word, is the representative of the corresponding picture
Create a database, put the visual word of each picture into the library, and build an index
Write CherryPy Web Apps, search for similar images based on clicked images

Extracting image feature points and generating Word files

Prerequisite to ensure that CMD executes the SIFT command
CMD executionStep1.py

#-*-Coding:utf-8-*-# step1.py: Extracts the feature points of a picture and generates a Word file vocabulary.pklImportPickleImportVocabularyImportImtoolsImportSift# imlist is a list of the name of the picture, the picture is placed under the static folderImlist = Imtools.get_imlist (' static/')# Total number of picturesNbr_images = Len (imlist)# Store the feature points of each picture in the corresponding. Sift signature fileFeatlist = [imlist[i][:-3]+' Sift '  forIinchRange (Nbr_images)] forIinchRange (nbr_images): Sift.process_image (Imlist[i], featlist[i])# Use K-means to train the corresponding words in the image feature file cluster# time relationship, only 46 images are used here, so only 46 words are createdVOC = vocabulary. Vocabulary (' Imagewords ') Voc.train (Featlist, $,Ten)# Save the words in Vocabulary.pkl withOpen' Vocabulary.pkl ',' WB ') asF:pickle.dump (VOC,F)# Print out the total number of wordsPrint ' vocabulary is: ', Voc.name, Voc.nbr_words

Store picture word information in a database

Database is SQLite
CMD execution:Step2.py

#-*-Coding:utf-8-*-# step2.py: According to the word file, the picture word into the SQLite databaseImportPickleImportSiftImportImagesearchImportImtools# List of picture namesImlist = Imlist = Imtools.get_imlist (' static/')# Number of picturesNbr_images = Len (imlist)# A list of image feature filesFeatlist = [imlist[i][:-3]+' Sift '  forIinchRange (Nbr_images)]# load Word file# put the word, picture name, address into the database images.db withOpen' Vocabulary.pkl ',' RB ') asF:VOC = Pickle.load (f) indx = Imagesearch. Indexer (' images.db ', VOC) indx.create_tables () forIinchRange (nbr_images): LOCS,DESCR = Sift.read_features_from_file (Featlist[i]) Indx.add_to_index (IMLIST[I],DESCR)# Submit the command to executeIndx.db_commit ()

Configure CherryPy Web Application

15 images randomly displayed when I click the image
Click on one of the cat pictures to search for similar pictures to show out
The essence is to look for words that are similar to the one in the database according to the words of the clicked image, and display them.
CMD executionStep3.py
Browse in Browser:127.0.0.1:8080

#-*-Coding:utf-8-*-# step3.py: Use CherryPy to do interactive interface, display the resultImportCherryPy, OS, urllib, pickleImportImtools fromNumPyImport*ImportImagesearch# CherryPy Page# Web root directory is set in configuration file service.conf# The default port is 8080 class searchimage:     def __init__(self):        # Load Picture Name listSelf.imlist = Imtools.get_imlist (' static/') self.nbr_images = Len (self.imlist) self.ndx = Range (self.nbr_images)# Load the generated Word filef = open (' Vocabulary.pkl ',' RB ') SELF.VOC = Pickle.load (f) f.close ()# Set the number of pictures to start showingSelf.maxres = the        # Set the structure of the pageSelf.header ="" " <!doctype html>  "" "Self.footer ="" </body>  "" "    # Response to index page    # Randomly display images when no search is found    # When searching for a picture similar to that of the picture, depending on the visual word     def index(self,query=none):SELF.SRC = Imagesearch. Searcher (' images.db ', SELF.VOC) HTML = self.header html + ="" " <br/> Click an image to search.            <a href= '? query= ' > Random selection </a> of images. <br/><br/> "" "        ifQuery# Pictures showing the results of the queryres = self.src.query (query) [: Self.maxres] forDist,ndxinchRes:imname = Self.src.get_filename (ndx) HTML + ="<a href= '? query="+imname+">"HTML + ="+imname+"' width= '/>"HTML + ="</a>"        Else:# randomly displayed picturesRandom.shuffle (SELF.NDX) forIinchSelf.ndx[:self.maxres]: Imname = self.imlist[i] html + ="<a href= '? query="+imname+">"HTML + ="+imname+"' width= '/>"HTML + ="</a>"HTML + = Self.footerreturnHTML index.exposed =True# Launch AppCherrypy.quickstart (Searchimage (),'/', Os.path.join (Os.path.dirname (__file__),' service.conf '))

Effect

When not searching:
When you click Search:

Similar applications

Wine clicks
Baidu Knowledge Map
Tutu Search Clothes

Search for similar images with images (visual words)

This article is an English version of an article which is originally in the Chinese language on aliyun.com and is provided for information purposes only. This website makes no representation or warranty of any kind, either expressed or implied, as to the accuracy, completeness ownership or reliability of the article or any translations thereof. If you have any concerns or complaints relating to the article, please send an email, providing a detailed description of the concern or complaint, to info-contact@alibabacloud.com. A staff member will contact you within 5 working days. Once verified, infringing content will be removed immediately.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

Get Started for Free

Sales Support

1 on 1 presale consultation

Chat Contact Sales
After-Sales Support

24/7 Technical Support 6 Free Tickets per Quarter Faster Response

Open a Ticket
Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.

Learn More