low level descriptors and high level descriptors

Types of visual descriptors[edit]

Descriptors are the first step to find out the connection between pixels contained in a digital image and what humans recall after having observed an image or a group of images after some minutes.

Visual descriptors are divided in two main groups:

1. General information descriptors: they contain low level descriptors which give a description about color, shape, regions, textures andmotion.
2. Specific domain information descriptors: they give information about objects and events in the scene. A concrete example would be face recognition.

General information descriptors[edit]

General information descriptors consist of a set of descriptors that covers different basic and elementary features like: color, texture, shape,motion, location and others. This description is automatically generated by means of signal processing.

• COLOR: the most basic quality of visual content. Five tools are defined to describe color. The three first tools represent the colordistribution and the last ones describe the color relation between sequences or group of images:
  • Dominant Color Descriptor (DCD)
  • Scalable Color Descriptor (SCD)
  • Color Structure Descriptor (CSD)
  • Color Layout Descriptor (CLD)
  • Group of frame (GoF) or Group-of-pictures (GoP)

• TEXTURE: also, an important quality in order to describe an image. The texture descriptors characterize image textures or regions. They observe the region homogeneity and the histograms of these region borders. The set of descriptors is formed by:
  • Homogeneous Texture Descriptor (HTD)
  • Texture Browsing Descriptor (TBD)
  • Edge Histogram Descriptor (EHD)

• SHAPE: contains important semantic information due to human’s ability to recognize objects through their shape. However, this information can only be extracted by means of a segmentation similar to the one that the human visual system implements. Nowadays, such a segmentation system is not available yet, however there exists a serial of algorithms which are considered to be a good approximation. These descriptors describe regions, contours and shapes for 2D images and for 3D volumes. The shape descriptors are the following ones:
  • Region-based Shape Descriptor (RSD)
  • Contour-based Shape Descriptor (CSD)
  • 3-D Shape Descriptor (3-D SD)

• MOTION: defined by four different descriptors which describe motion in video sequence. Motion is related to the objects motion in the sequence and to the camera motion. This last information is provided by the capture device, whereas the rest is implemented by means of image processing. The descriptor set is the following one:
  • Motion Activity Descriptor (MAD)
  • Camera Motion Descriptor (CMD)
  • Motion Trajectory Descriptor (MTD)
  • Warping and Parametric Motion Descriptor (WMD and PMD)

• LOCATION: elements location in the image is used to describe elements in the spatial domain. In addition, elements can also be located in the temporal domain:
  • Region Locator Descriptor (RLD)
  • Spatio Temporal Locator Descriptor (STLD)

Specific domain information descriptors[edit]

These descriptors, which give information about objects and events in the scene, are not easily extractable, even more when the extraction is to be automatically done. Nevertheless they can be manually processed.

As mentioned before, face recognition is a concrete example of an application that tries to automatically obtain this information.