Skip to main content Skip to first level navigation

RICOH imagine. change.

Ricoh Software Research Center(Beijing) Co., Ltd.

Skip to main content First level navigation Menu
ricoh.com
Breadcrumbs

R&D Activities

Visual Processing Technology

Directions

  • Object Detection and Recognition
  • Image Processing and Analysis
  • Human–Computer Interaction

Applications

Multi-pattern Auto White Balance Technology

For scenes with different light sources, traditional white balance method of digital camera could not set up optimal parameters for all different regions. SRCB proposed a multi-pattern auto white balance method based on image recognition technology. The method can define white balance to fit the light source of each segregated area in the image. This new capability is very powerful in improving photo quality. The multi-pattern auto white balance method has been adopted to Ricoh digital camera successfully.

Dynamic Range Double Shot Technology

It can be difficult to photograph scenes where the level of brightness varies greatly. To extend dynamic range of Ricoh digital camera, and to improve photo quality, SRCB designed dynamic range double shot method based on image fusion technique. The proposed algorithm could expand the dynamic range up to a maximum equivalent of 12 EV (makes it possible to record images similar to what naked-eye can see). The technology has been deployed in Ricoh digital camera successfully.

Image Auto Annotation Technique

The goal of image auto annotation technique is to automatically analyze and extract the contents of an image using Computer Vision techniques, and to annotate the image with keywords relating to the extracted topics. SRCB proposed a new EGCap algorithm based on graph-based captioning method. The performance of EGCap algorithm achieved the highest level all over the world and a related paper was published in the proceeding of ACM-MIR2008.

Research on vehicle application based on new generation camera

  • Image preprocessing
  • Environment recognition
  • Object segmentation
  • Vehicle detection & Tracking
  • Pedestrian detection
  • Object detection & Tracking

R&D of human computer interface based on gesture recognition

  • Object detection and localization based on RGB-D camera
  • Gesture recognition based on RGB-D camera
  • Application of gesture recognition(Tele meeting, etc.)