Language and Script Identification Based on Steerable Pyramid Features

Abstract

Arabic and Latin language and script identification in machine printed and handwritten types present several difficulties because the Arabic (machine printed or handwritten) and the handwritten Latin scripts are cursive scripts of nature. To avoid all possible confusions which can be generated, we propose in this paper an accurate and suitable designed system for language and script identification at word level which is based on steerable pyramid transform. The features extracted from pyramid sub bands serve to classify the scripts on only one script among the scripts to identify. The encouraging and promising results obtained are presented in this research paper.

Arabic/Latin and Machine-printed/Handwritten Word Discrimination using HOG-based Shape Descriptor

“…We compared CP-HOG (our proposed system) with systems proposed by [7] and [28] which deal with the same problem. To compare CP-HOG with the steerable pyramid transform, proposed in [7] work, we used a common database (our database composed of 4000 word samples).…”

“…A/L P (1976T, 8320W) T/W 99.7%(T), 96.8%(W) [7] A/L P/H (800W) W 97.5% [8] A Zheng et al [4] proposed features from Run-length histogram for machine-printed/handwritten Chinese character classification. Black pixel run-lengths are extracted in three directions: horizontal, vertical and diagonal.…”

“…Horizontal H, vertical V , major diagonal M D and minor diagonal mD orientations and 2 pixels distance d level for the classification are used (see Table 3). Authors in [7] proposed a script identification system, at word level, based on steerable pyramid transform. A steerable pyramid decomposition is a linear multi-orientation, multi-resolution image decomposition method by which an image is subdivided into a collection of sub-bands localized at different scales and orientations.…”

“…Most of used features are statistical such as connected component width, height, aspect ratio, area and density and separator length between two successive connected components [13], connected component profiles analysis [10], loop aspect ratio [1], pixel distribution and upper lower profile by [3], baseline profile features, word physical sizes, overlapping areas, crossing count histogram [2], bi-level co-occurrence [9], run-length histogram [4], Gabor filters [6] and [38], steerable pyramids transform [7] or structural (ascenders, descenders, loops, diacritic signs, etc. ).…”

Density Based Script Identification of a Multilingual Document Image

“…Many techniques have been proposed to identify various scripts. In recent advances, the scripts are identified using morphological reconstruction & hole filling [8], water reservoir techniques [9], projection profiles [10], moments & run-length histogram [3], top & bottom profiles [11] [12], vertical & horizontal profiles [13], steerable pyramid features [14], texture features [15] & wavelet energy histogram. Entropy or energy evaluation has been done in a multilingual document image wherein column entropies are considered for the purpose of identification [16].…”