Face Tracer: A Search Engine for Large Collections of Images with Faces

Face Tracer: A Search Engine for Large Collections of Images with Faces

The face tracer engine has been created and the first image search engine would contain images based only on the faces. In this search engine one can find the required image by simple entering the queries like smiling girls, men with blonde hair and man with mustaches. In each search the probability of the search results is the 3.1 million faces that would come by entering the labels of the different face attributes.

Faces that have been provided in the face tracer engine are also aligned with the images from the internet, form the commercial use, from different face detectors etc. moreover, these face have also been labeled by their respective face attributes. Ina addition to this the face tracer systems can be used in a variety of applications and by a number of different departments. Such as, social networks, photo management departments, law enforcement departments. This search engine will be readily and soon be available in the market for the sale.

This is the first face tracer engine and it allows its users and the operators to search through a large number of image collections. Besides that this collection is continuously being enhanced.

This system let its workers and users to search variety of facial attributes using natural and casual language queries small children, women with blue eyes, bald men etc.

Most of the previously mention face search engines have limited search abilities and do not have larger collections. They show only limited search results on the basis of the annotation of the entered text and words.  But, there are many issues in the search images which are based on the annotations. Such as, the manual classification of images is time taking and consuming. Moreover, the annotations can be incorrect due to spelling mistakes sometimes they can be misleading and not rightly entered.

The change in quality and the alteration of the search results is obviously visible. Google search engine relies in the annotations of the text and most of the time sits search results are not relevant to the queries and the text that has been asked. But the face tracer system is more efficient and reliable. It can returns and depicts only those images that will match the query.

This powerful approach has been developed and obtained by developing and labeling the images and by marking the facial attributes on them. So the search results will be come up and the automatically labeling images off-line on the basis of a large number of attributes. At images labeled will match the entered text will show up.

Moreover, it is also quite easy and facilitating to improve and add new images and more face attributes to our face tracer search engine, allowing for forthcoming scalability. Signifying new characteristics and labeling by hand faces to match those attributes can also be done collaboratively and by a combined effort by a community of users of that search engine

In this research work we had developed and describe a new face tracer engine and this new approach can search for the images in bulky databases and we have also constructed and developed the first face search engine using this approach. This search engine is limited in scope as it will only show the and focused on the face images of the people and will bring into line the images to a common coordinate system.

As this system would be more established and continue to grow some improvement would be made in the system and some of the current limitations will also be addressed and improved.

For instance, the system will be developed to provide more than just images and 2D and 3D images will also be provided. Besides that region specifying characteristics of faces will also be provided in the search.