Menu Content/Inhalt
Computer Vision and Digital Signal Processing Group (CVDSP-UJI) PDF Print E-mail

The CVDSP-UJI research tem is comprised of researchers (14 Ph. D. and 8 Ph. D. students) from the following research groups:

The  CVDSP-UJI research group is structured in three sections. The research interests of the sections lie in the following areas:

Automatic Learning and Pattern Analysis section

  • Classifiers based on distance, decision trees, Bayesian classifiers, support vector machines.
  • Information reduction techniques: selection and feature extraction, prototypes selection, features-prototypes hybrid methods.
  • Combination of classifiers: boosting, bagging. 
  • Scalability of the classification techniques: set of samples partitions, problems with unbalanced distributions of classes. 
  • Semisupervised and continuous learning, adaptive methods, fuzzy logic, clustering.

Image Processing section

  • Color and multispectral image processing. 
  • Image features extraction. 
  • Supervised and unsupervised segmentation.
  • Texture analysis. 
  • Image filtering. 
  • Recovery of visual information in databases. 
  • Visual pattern recognition.
Video analysis section

  • Mono and multimodal image registration.
  • Stereoscopic images analysis and 3D reconstruction. 
  • Motion detection, estimation and segmentation. 
  • Tracking and active vision 
  • Actions recognition in video sequences and motion characterization. 
  • Annotation and classification of video sequences

Application fields

  • Applications on large volumes of visual data. 
  • Quality control and visual inspection in industrial production. 
  • Aerial and satellite imagery. 
  • Biomedical applications. 
  • Advanced interfaces based on computer vision. 
  • Surveillance systems, Recognition and motion monitoring for people and vehicles. 
  • Cognitive systems. Learning systems and adaptation to the environment based on vision.