Overview of the MPEG-CDVS Standard.

Area of Science:

• Computer Vision
• Image Processing
• Information Retrieval

Background:

• The need for efficient image retrieval systems is growing.
• Interoperability in visual search applications is crucial.
• Existing methods for image feature extraction can be computationally intensive and require large storage.

Purpose of the Study:

• To introduce the ISO/IEC moving pictures experts group (MPEG) Compact Descriptors for Visual Search (CDVS) standard.
• To detail the technical features and evolution of the MPEG-CDVS standard.
• To highlight improvements in data size and computational efficiency for image retrieval.

Main Methods:

• Standardization of bitstream syntax for image feature data.
• Development of efficient feature extraction algorithms.
• Optimization of computation and memory footprint.

Main Results:

• Achieved significant reductions in image feature data size.
• Demonstrated reduced computation and memory requirements for feature extraction.
• Established a standardized bitstream syntax for interoperable visual search.

Conclusions:

• The MPEG-CDVS standard offers a robust solution for efficient and interoperable image retrieval.
• The standard facilitates practical implementation of visual search applications.
• Continued evolution of CDVS promises further advancements in image search technology.