Pattern Recognition by Data Compression
نویسنده
چکیده
Combinatorialpattern recognition is used for discrimination of handwritten signatures. The method partitions an input image into overlapping pieces. A piece is as large as possible under the constraints rst that it must have an identical copy at a smaller row number, second that the piece must have a xed width. The partitioning into pieces is performed according to a generalisation of the Lempel{Ziv procedure for lossless data compressionn2]. A set of programs has been implemented. Applications are shown for recognition of graphical patterns in printed text and for discrimination of handwritten signatures.
منابع مشابه
On the Contribution of Compression to Visual Pattern Recognition
Most pattern recognition problems are solved by highly task specific algorithms. However, all recognition and classification architectures are related in at least one aspect: They rely on compressed representations of the input. It is therefore an interesting question how much compression itself contributes to the pattern recognition process. The question has been answered by Benedetto et al. (...
متن کاملData Compression - A Generic Principle of Pattern Recognition?
Most pattern recognition problems are solved by highly task specific algorithms. However, all recognition and classification architectures are related in at least one aspect: They rely on compressed representations of the input. It is therefore an interesting question how much compression itself contributes to the pattern recognition process. The question has been answered by Benedetto et al. (...
متن کاملInfluence of compression on 3D face recognition
This paper studies how the performance of a 3D face recognition system is affected by compression. A novel lossy compression technique tailored for registered 3D data along with a scheme for 3D face registration and recognition are presented and the results discussed. The proposed scheme achieves a significant compression ratio (factor of 35) without the loss of recognition performance. 2009 El...
متن کاملStandards-Compatible Compression for Automated Image Recognition in Sensor Networks and Surveillance Systems
We show that the standard image compression algorithms are not suitable for compressing images in correlation pattern recognition since they aim at retaining image fidelity in terms of perceptual quality rather than preserving spectrally significant information for pattern recognition. New compression algorithms for pattern recognition are therefore developed, which are based on the modificatio...
متن کاملSimultaneous Secure Transmission for Textual Data Using Radomized Bits
Security of data transmitted through internet has put forward a number of challenges. In this thesis work, system is developed in which two techniques are combined, encryption and compression, which provides a strong backbone for its security and reduces extra overhead. The proposed scheme extends security by incorporating pattern recognition, data encryption using SDES and XOR encryption techn...
متن کاملVector Quantization Parallelization
Quantization is the process of representing a large set of input values with a much smaller set. In signal processing and image processing, Vector Quantization is a classical quantization which extends the scalar quantization to multi-dimensional space. It is widely used in many applications such as data compression, data correction, pattern recognition, and density estimation. This project pro...
متن کامل