Comparing String Representations and Distances in a Natural Images Classification Task
نویسندگان
چکیده
This paper shows how strings can be used in a natural images classification task. We propose to build an attributed string from a set of regions of interest detected thanks to an interest point detector. These salient zones are characterized by local signatures describing singularities and they are linked by using graph seriation algorithms and perceptual methods. Once each image is represented by a string of signatures, we propose to use string-based edit distances and an ordered histogramsbased distance in order to perform the classification task. Experiments have shown that whereas seriation algorithms give approximately the same results, the ordered histogram based distance is more efficient for the considered application.
منابع مشابه
پتانسیلهای محض SU(3) در مدل ورتکسهای مرکزی پهن
The potentials between static sources in various representations in SU(3) are calculated on the basis of the fat-centre-vortices model of Faber, Greensite and Olejník. At intermediate distances, potentials are in qualitative agreement with “Casimir scaling,” which says that the string tension is proportional to the quadratic operator of the representation. At large distances, screening occurs...
متن کاملWeighted Symbols-Based Edit Distance for String-Structured Image Classification
As an alternative to vector representations, a recent trend in image classification suggests to integrate additional structural information in the description of images in order to enhance classification accuracy. Rather than being represented in a p-dimensional space, images can typically be encoded in the form of strings, trees or graphs and are usually compared either by computing suited met...
متن کاملImprovement of the Classification of Hyperspectral images by Applying a Novel Method for Estimating Reference Reflectance Spectra
Hyperspectral image containing high spectral information has a large number of narrow spectral bands over a continuous spectral range. This allows the identification and recognition of materials and objects based on the comparison of the spectral reflectance of each of them in different wavelengths. Hence, hyperspectral image in the generation of land cover maps can be very efficient. In the hy...
متن کاملComparing the Capability of Sentinel 2 and Landsat 8 Satellite Imagery in Land Use and Land Cover Mapping Using Pixel-based and Object-based Classification Methods
Introduction: Having accurate and up-to-date information on the status of land use and land cover change is a key point to protecting natural resources, sustainable agriculture management and urban development. Preparing the land cover and land use maps with traditional methods is usually time and cost consuming. Nowadays satellite imagery provides the possibility to prepare these maps in less ...
متن کاملLearning Document Image Features With SqueezeNet Convolutional Neural Network
The classification of various document images is considered an important step towards building a modern digital library or office automation system. Convolutional Neural Network (CNN) classifiers trained with backpropagation are considered to be the current state of the art model for this task. However, there are two major drawbacks for these classifiers: the huge computational power demand for...
متن کامل