GazoPa is an image search engine that launched recently on the Web. With most of the popular search engines such as Google, Baidu, Yahoo!, Microsoft’s Live Search have incorporated image search as one of the search services rendered, GazoPa is a visual image search engine that solely generating results for similar images.
GazoPa, currently in invitation-only beta phase, it enables users to perform similar image search for more than 50 million images on the Web through its proprietary technology. On the GazoPa site, you can search images by uploading your own image file, drawing, or type a keyword, and even if you do not have any inventory of images, you can type a URL link to GazoPa’s search bar. GazoPa will then generating all the similar images based on the characteristics of color and shade extracted from the image itself. The results generated might not in the form of images, it can also be appeared as videos, as GazoPa supports video thumbnails search by using image.
When testing the accuracy of GazoPa’s image search service, I’ve uploaded a satin ribbon photo. Not only it did not generate results based on the same category, it also did not render my desired results, i.e. a bunch of products (See the picture shown in below). Based on the search results, perhaps I would be understood GazoPa more if I knew that GazoPa’s image results are determined on the similarity of images instead of the sameness of images in the first place. On contrary, I performed an image search with the same item using the aforesaid TinEye, zero results appeared at the time of my writing. However, as I used GazoPa, it will definitely generate all the images available in its database based on the similar color, in this case the images with the light gray color.
In using GazoPa, users are allowed to search through playing tools such as trimming, region fill, or eye dropper tool that act as filters. There is also a search option in which you can click any one of the parameters, such as color, moderate1, moderate 2, shape, face, image size, monochrome only, or omit some images, thus you’ll get to learn the different search aspects rendered by GazoPa.
At this moment, GazoPa seem to be able to search images available in MySpace, YouTube, Twitter, Wikipedia, Yahoo!, Alibaba, but my general sense is that its actual search will need to be more robust to be most useful.
To know more about GazoPa, it is headquartered in Japan and being backed by Hitachi.
