HP Labs Multimedia Analytic Platform (beta)
Photos of people represent a large majority of all the images taken. This is especially
true for the social networking websites. The Face Detection service detects all
the faces in a given image and returns their bounding boxes.
The Face Verification service is a continuation of the Face Detection service in
that faces from different images are compared and their similarity is computed.
Face demographics estimation provides the age and gender of the faces in an image. It is a basic building block for many high level applications such as relationship estimation, audience measurement. Our demographics estimation service returns the demographics information for each face in a given image.
In many imaging applications, an image is represented as feature vectors so that
it can be compared and matched to other images. In this service, highly efficient
and compact features are extracted to represent an image. The returned features include
feature locations in the image and their vector descriptions.
In many imaging applications, an image needs to be matched to another image to compute
similarity or detect near-duplications. This cloud service returns a similarity
score for two images based on the feature matching.
Image collage service is supported by PhotoShape technology, which enables the automatic creation of compact and informative image collages of arbitrary geometrical shapes with arbitrary numbers of images. Notably, this technology is capable of generating image collages of complex shapes with small numbers of images (even a single image), which is not accomplished by existing competitive technologies.
When we take pictures we often don't have time to compose them carefully. In addition,
the fixed shape rectangle of our camera's eye is often not ideally suited to the
subject we are taking. Cropping an image can help solve both these problems. AUTO
CROP takes the tedium out of manual cropping by automatically suggesting an improved
composition for many of your images. It uses advanced image understanding algorithms
coupled with an appreciation of what makes a picture look great.
Time Cluster will split the set of photograhps to some cluster by time.
Duplicate Cluster will split the set of photograhps to some clusters. if two images
are similar, then they will be put in the the same cluster.
Sharpness will analyzed the photograhp and returns a value of sharpness.