Several content-based image retrieval systems

Content-Based Image Retrieval System

Reprinted from: http://blog.csdn.net/flyingworm_eley/article/details/6521765

Certificate ------------------------------------------------------------------------------------------------------------------------------------------

Content-based image retrieval has been developed for more than 20 years, and basic and simple technologies have matured. Let's summarize several systems.

The following ten search engines can help you find images, search for images, and search for similar images.

1: http://tineye.com/

Tineye is a typical search engine for image search. You can enter images on the local hard disk or the image URL to automatically search for similar images. The search accuracy is relatively satisfactory.

Tineye is a similar image search engine developed by Canadian company id é e. Its main purposes include: 1. Discovering image sources and related information; 2. Researching and tracing the dissemination of Image Information on the Internet; 3. Find the image of the high-resolution version; 4. Find the webpage with your photo; 5. Check the different versions of the image.

2. http://shitu.baidu.com

Baidu officially launched its latest search function, "Map Recognition" (shitu.baidu.com ). This function is based on similar image recognition technology. After users upload a local image or enter the URL of the image, Baidu analyzes the Image Based on the image features, then, you can search for similar image resources and information from the Internet. However, when you upload a local image, the image file size is smaller than 5 MB. The format can be JPG, JPEG, GIF, PNG, BMP, and other image files.

Http://www.gazopa.com/

Gazopa does not search for Images Based on keywords, but searches for images based on certain features of the image, such as color and shape. There are four gazopa search methods:

The first is the traditional keyword search for images, but gazopa cannot compete with search engines such as Google in the traditional image search field.

The second is innovative image search, but gazopa cannot compete with tineye in this field. Tineye can easily search for the results closest to the source image, while gazopa often does not match the source image.

The third method is to search for images by hand-drawn images. This method is not very useful. Although gazopa has such shortcomings, it is also a very original search engine. Gazopa is still in the internal test phase. If you want to join the test, you can leave your email address on the official website. After receiving the invitation, you can test it.

 

4. http://similar-images.googlelabs.com/

Google lab Image Search: after entering a keyword, for example, "Lake", click similar images under an image on the returned page, and use the image search engine similar to Google, you can search for all the similar images and display them to users for viewing. Its Accuracy and accuracy are relatively high.

5. http://www.picitup.com/

Picitup is a professional image search engine in Public Beta. It has powerful functions and supports searching for Chinese keywords. It is a good choice for image lovers in China. Picitup mainly supports keyword search, but in its special search item celebritymatchup, you can upload local photos for search, but the results are generally disappointing. Picitup can select a filtering method on the search results page to filter images. For example, you can filter the search results by color, Avatar (FACE), landscape, or product.

The biggest feature of picitup is to provide similar image search, that is, to find the initial image through the keyword, click the similar Pictures button under the initial image to search for images similar to this image. The essence is the same as that of Google lab image search.

6. http://www.tiltomo.com/

Tiltomo is a search tool developed by Flickr. It is mainly used to maintain the image database of Flickr. Its search algorithm is mainly based on similar theme styles or similar colors and materials.

7. http://cn.bing.com/

Live.com allows you to perform a keyword search and then perform a similarity search. You can search for similar images for any image in the live index, but the search results do not seem very accurate.
8. http://www.xcavator.net

Xcavator and live.com are very similar. You need to enter a keyword, select an image from the search results, and search for the image based on its characteristics.

9. http://www.incogna.com

The search speed of incogna is very fast, mainly based on the similarity between colors and shapes.

10. http://www.terragalleria.com

Terragalleria is mainly based on visual similarity, regardless of the image content.

11 http://labs.ideeinc.com/upload/
BYO image search searches for similar images based on the images you upload. The algorithms are mainly based on colors and themes.

12, Amoy search: http://www.taotaosou.com/