Arabic Digits Research Articles

Arabic digits speech recognition and speaker identification in noisy environment using a hybrid model of VQ and GMM

This paper presents an automatic speaker identification and speech recognition for Arabic digits in noisy environment. In this work, the proposed system is able to identify the speaker after saving his voice in the database and adding noise. The mel frequency cepstral coefficients (MFCC) is the best approach used in building a program in the Matlab platform; also, the quantization is used for generating the codebooks. The Gaussian mixture modelling (GMM) algorithms are used to generate template, feature-matching purpose. In this paper, we have proposed a system based on MFCC-GMM and MFCC-VQ Approaches on the one hand and by using the Hybrid Approach MFCC-VQ-GMM on the other hand for speaker modeling. The White Gaussian noise is added to the clean speech at several signal-to-noise ratio (SNR) levels to test the system in a noisy environment. The proposed system gives good results in recognition rate.

Shared Numerosity Representations Across Formats and Tasks Revealed with 7 Tesla fMRI: Decoding, Generalization, and Individual Differences in Behavior.

Debate continues on whether encoding of symbolic number is grounded in nonsymbolic numerical magnitudes. Nevertheless, fluency of perceiving both number formats, and translating between them, predicts math skills across the life span. Therefore, this study asked if numbers share cortical activation patterns across formats and tasks, and whether neural response to number predicts math-related behaviors. We analyzed patterns of neural activation using 7 Tesla functional magnetic resonance imaging in a sample of 39 healthy adults. Discrimination was successful between numerosities 2, 4, 6, and 8 dots and generalized to activation patterns of the same numerosities represented as Arabic digits in the bilateral parietal lobes and left inferior frontal gyrus (IFG) (and vice versa). This indicates that numerosity-specific neural resources are shared between formats. Generalization was also successful across tasks where participants either identified or compared numerosities in bilateral parietal lobes and IFG. Individual differences in decoding did not relate to performance on a number comparison task completed outside of the scanner, but generalization between formats and across tasks negatively related to math achievement in the parietal lobes. Together, these findings suggest that individual differences in representational specificity within format and task contexts relate to mathematical expertise.

Arabic digit recognition using robust deep convolution neural network

Recently, digit recognition becomes one of interest problems for many researchers. However, Arabic digits have lack for such research. In this work, we used a robust deep convolution neural network (DCNN) to evaluate our collected Arabic digit dataset. We introduce substantial changes to CNN models to achieve superior results to prior work. Extensive experimental results are contacted to show the robustness of proposed models. Our detector achieves best current state-of-the-art results.

FEATURE EXTRACTION ALGORITHM USING NEW CEPSTRAL TECHNIQUES FOR ROBUST SPEECH RECOGNITION

In this work, we propose a novel feature extraction algorithm that improves the robustness of automatic speech recognition (ASR) systems in the presence of various types of noise. The proposed algorithm uses a new cepstral technique based on the differential power spectrum (DPS) instead of the power spectrum (PS), the algorithm replaces the logarithmic non linearity by the power function. In order to reduce cepstral coefficients mismatches between training and testing conditions, we used the mean and variance normalization, then we apply auto-regression movingaverage filtering (MVA) in the cepstral domain. The ASR experiments were conducted using two databases, the first is LASA digit database designed for recognition the isolated Arabic digits in the presence of different types of noise. The second is Aurora 2 noisy speech database designed to recognize connected English digits in various operating environments. The experimental results show a substantial improvement from the proposed algorithm over the baseline Mel Frequency Cepstral Coefficients (MFCC), the relative improvement is the 28.92% for LASA database and is the 44.43% for aurora 2 database. The performance of our proposed algorithm was tested and verified by extensive comparisons with the state-of-the-art noise-robust features in aurora 2.

Automaticity in processing spatial-numerical associations: Evidence from a perceptual orientation judgment task of Arabic digits in frames.

Human adults are faster to respond to small/large numerals with their left/right hand when they judge the parity of numerals, which is known as the SNARC (spatial-numerical association of response codes) effect. It has been proposed that the size of the SNARC effect depends on response latencies. The current study introduced a perceptual orientation task, where participants were asked to judge the orientation of a digit or a frame surrounding the digit. The present study first confirmed the SNARC effect with native Chinese speakers (Experiment 1) using a parity task, and then examined whether the emergence and size of the SNARC effect depended on the response latencies (Experiments 2, 3, and 4) using a perceptual orientation judgment task. Our results suggested that (a) the automatic processing of response-related numerical-spatial information occurred with Chinese-speaking participants in the parity task; (b) the SNARC effect was also found when the task did not require semantic access; and (c) the size of the effect depended on the processing speed of the task-relevant dimension. Finally, we proposed an underlying mechanism to explain the SNARC effect in the perceptual orientation judgment task.

The perceived duration of numerical and verbal digits: The independent effects of digit value and covered area.

Numerical digits influence perceived duration; specifically, higher numerical digit values are perceived to be longer than lower ones that are presented for the same duration of time. To examine the functional relationship between digit value and perceived duration further, three models based on digit value and covered area were tested in two experiments. Unlike previous studies, each experiment utilised the entire digit range that spans from 1 to 9. In both experiments, digit values were presented visually during the target interval of a time reproduction task. Although numerical digits (i.e., Arabic digits) were used in Experiment 1, verbal digits (i.e., number words) were utilised in Experiment 2. In the case of the numerical digits, perceived duration increased as a function of digit value, while the effect of covered area failed to reach statistical significance. For verbal digits, however, the effects of both digit value and covered area were statistically significant. Overall, our findings support the existence of two distinct nontemporal processes that are involved in the modulation of perceived duration: one process that is based on a digit's surface characteristics (i.e., covered area) and another process that is based on the digit's numerical meaning (i.e., digit value).

Improving the Accuracy for Offline Arabic Digit Recognition Using Sliding Window Approach

Handwritten Digit recognition is a challenging problem these days due to the widely used Arabic language in the world, especially in the Middle East region. In this paper, sliding windows are used to enhance classification accuracies and implemented using random forests (RF) and support vector machine (SVM) classifiers for recognition of Arabic digit images. In order to study their effectiveness with and without using sliding windows, four different feature extraction techniques have been proposed which includes Mean-based, Gray-Level Co-occurrence Matrix (GLCM), Moment-based, and Edge Direction Histogram (EDH). The obtained accuracies show the significance of using sliding windows for classifying digit. The recognition rates acquired using the modified version of AHDBase dataset are 98% when Mean-based and Moment-based are applied with RF classifier, 98.33% and 99.13% when GLCM and EDH are used with linear-kernel SVM, respectively. Moreover, the performance of this study is compared against recent state-of-the-art approaches, namely Geometric-based, two-dimensional discrete cosine transform, Hierarchical features, Hetero-features, Discrete Fourier Transform and geometrical features, Gabor-based, gradient, structural, and concavity and Local Binary Convolutional Neural Networks.

Network topology of symbolic and nonsymbolic number comparison.

Studies of brain activity during number processing suggest symbolic and nonsymbolic numerical stimuli (e.g., Arabic digits and dot arrays) engage both shared and distinct neural mechanisms. However, the extent to which number format influences large-scale functional network organization is unknown. In this study, using 7 Tesla MRI, we adopted a network neuroscience approach to characterize the whole-brain functional architecture supporting symbolic and nonsymbolic number comparison in 33 adults. Results showed the degree of global modularity was similar for both formats. The symbolic format, however, elicited stronger community membership among auditory regions, whereas for nonsymbolic, stronger membership was observed within and between cingulo-opercular/salience network and basal ganglia communities. The right posterior inferior temporal gyrus, left intraparietal sulcus, and two regions in the right ventromedial occipital cortex demonstrated robust differences between formats in terms of their community membership, supporting prior findings that these areas are differentially engaged based on number format. Furthermore, a unified fronto-parietal/dorsal attention community in the nonsymbolic condition was fractionated into two components in the symbolic condition. Taken together, these results reveal a pattern of overlapping and distinct network architectures for symbolic and nonsymbolic number processing.

Verbal working memory load dissociates common indices of the numerical distance effect: Implications for the study of numerical cognition

In four experiments, we explore the role that verbal WM plays in numerical comparison. Experiment 1 demonstrates that verbal WM load differentially impacts the two most common variants of numerical comparison tasks, evidenced by distinct modulation of the size of the numerical distance effect (NDE). Specifically, when comparing one Arabic digit to a standard, the size of the NDE increases as a function of increased verbal WM load; however, when comparing two simultaneously presented Arabic digits, the size of the NDE decreases (and here is eliminated) as a function of an increased verbal WM load. Experiment 2, using the same task structure but different stimuli (physical size judgments), provides support for the notion that this pattern of results is unique to tasks employing numerical stimuli. Experiment 3 demonstrates that the patterns observed in Experiment 1 are not an artifact of the stimulus pairs used. Experiment 4 provides evidence that the differing pattern of results observed between Experiment 1 and Experiment 2 are due to differences in stimuli (numerical vs. non-numerical) rather than to other differences between tasks. These results are discussed in terms of current theories of numerical comparison.

Is the long-term association between symbolic numerical magnitude processing and arithmetic bi-directional?

By analyzing longitudinal data from the start to the end of primary education, we aimed to investigate whether symbolic numerical magnitude processing at the start of primary education predicted arithmetic at the end, and whether arithmetic at the start of primary education predicted later symbolic numerical magnitude processing skills at the end. In the first grade (start) and sixth grade (end) of primary education, the same group of children’s symbolic numerical magnitude processing skills and arithmetic competence were assessed. We were particularly interested in exploring the direction of the association between symbolic numerical magnitude processing and arithmetic and observed that this association was bi-directional across primary education. Symbolic numerical magnitude processing skills in first grade predicted arithmetic in sixth grade; but also the reversed direction turned out significant: Early arithmetic predicted later symbolic numerical magnitude processing skills. Both directions remained significant after controlling for motor speed and nonverbal reasoning. Critically, when controlling for auto-regressive effects of prior abilities, the symbolic comparison-arithmetic association was no longer significant, the reversed direction became marginally significant. This suggests that children’s arithmetic development across primary education to some extent strengthens their ability to process the numerical meaning of Arabic digits.

A shared code for Braille and Arabic digits revealed by cross-modal priming in sighted Braille readers

Quantities can be represented by different formats (e.g. symbolic or non-symbolic) and conveyed via different modalities (e.g. tactile or visual). Despite different priming curves: V-shape and step-shape for place and summation coded representation, respectively, the occurrence of priming effect supports the notion of different format overlap on the same mental number line. However, little is known about tactile-visual overlap of symbolic numerosities i.e. Braille numbers to Arabic digits on the magnitude number representation. Here, in a priming experiment, we tested a unique group of sighted Braille readers to investigate whether tactile Braille digits would activate a place-coding type of mental number representation (V-shape), analogous to other symbolic formats. The primes were either tactile Braille digits presented on a Braille display or number words presented on a computer screen. The targets were visually presented Arabic digits, and subjects performed a naming task. Our results reveal a V-shape priming function for both prime formats: tactile Braille and written words representing numbers, with strongest priming for primes of identical value (e.g. “four” and “4”), and a symmetrical decrease of priming strength for neighboring numbers, which indicates that the observed priming is due to identity priming. We thus argue that the magnitude information is processed according to a shared phonological code, independent of the input modality.

A comes before B, like 1 comes before 2. Is the parietal cortex sensitive to ordinal relationships in both numbers and letters? An fMRI-adaptation study.

How are number symbols (e.g., Arabic digits) represented in the brain? Functional resonance imaging adaptation (fMRI‐A) research has indicated that the intraparietal sulcus (IPS) exhibits a decrease in activation with the repeated presentation of the same number, that is followed by a rebound effect with the presentation of a new number. This rebound effect is modulated by the numerical ratio or difference between presented numbers. It has been suggested that this ratio‐dependent rebound effect is reflective of a link between the symbolic numerical representation system and an approximate magnitude system. Experiment 1 used fMRI‐A to investigate an alternative hypothesis: that the rebound effect observed in the IPS is related to the ordinal relationships between symbols (e.g., 3 comes before 4; C after B). In Experiment 1, adult participants exhibited the predicted distance‐dependent parametric rebound effect bilaterally in the IPS for number symbols during a number adaptation task, however, the same effect was not found anywhere in the brain in response to letters. When numbers were contrasted with letters (numbers > letters), the left intraparietal lobule remained significant. Experiment 2 demonstrated that letter stimuli used in Experiment 1 generated a behavioral distance effect during an active ordinality task, despite the lack of a neural distance effect using fMRI‐A. The current study does not support the hypothesis that general ordinal mechanisms underpin the neural parametric recovery effect in the IPS in response to number symbols. Additional research is needed to further our understanding of mechanisms underlying symbolic numerical representation in the brain.

Inversion effects on mental arithmetic in English- and Polish-speaking adults

In some languages the order of tens and units in number words is inverted compared with the symbolic digital notation (e.g., German 23 → "dreiundzwanzig," literally: "three-and-twenty"). In other languages only teen-numbers are inverted (e.g., English 17 → "seventeen"; Polish 17 → "siedemnaście" literally "seventeen"). Previous studies have focused on between group comparisons of inverted and non-inverted languages and showed that number word inversion impairs performance on basic numerical tasks and arithmetic. In two independent experiments, we investigated whether number word inversion affects addition performance within otherwise non-inverted languages (Exp. 1: English, Exp. 2: Polish). In particular, we focused on the influence of inverted (I; English: teen-numbers ⩾ 13, Polish: numbers 11-19) and non-inverted (N) summands with sums between 13 and 39. Accordingly, three categories of addition problems were created: N + N, N + I, and I + I with problem size matched across categories. Across both language groups, we observed that problems with results in the 20 and 30 number range were responded to faster when only non-inverted summands were part of the problems as opposed to problems with one or two inverted summands. In line with this, the cost of a carry procedure was the largest for two inverted summands. The results support the notion that both language-specific and language-invariant aspects contribute to addition problem-solving. In particular though, regarding language-specific aspects, the results indicate that inverted number word formation of teens influences place-value processing of Arabic digits even in otherwise non-inverted languages.

Making Sense of Number Words and Arabic Digits: Does Order Count More?

The ability to choose the larger between two numbers reflects a mature understanding of the magnitude associated with numerical symbols. The present study explores how the knowledge of the number sequence and memory capacity (verbal and visuospatial) relate to number comparison skills while controlling for cardinal knowledge. Preschool children's (N=140, Mage-in-months =58.9, range=41-75) knowledge of the directional property of the counting list as well as the spatial mapping of digits on the visual line were assessed. The ability to order digits on the visual line mediated the relation between memory capacity and number comparison skills while controlling for cardinal knowledge. Beyond cardinality, the knowledge of the (spatial) order of numbers marks the understanding of the magnitude associated with numbers.

Challenging the neurobiological link between number sense and symbolic numerical abilities.

A significant body of research links individual differences in symbolic numerical abilities, such as arithmetic, to number sense, the neurobiological system used to approximate and manipulate quantities without language or symbols. However, recent findings from cognitive neuroscience challenge this influential theory. Our current review presents an overview of evidence for the number sense account of symbolic numerical abilities and then reviews recent studies that challenge this account, organized around the following four assertions. (1) There is no number sense as traditionally conceived. (2) Neural substrates of number sense are more widely distributed than common consensus asserts, complicating the neurobiological evidence linking number sense to numerical abilities. (3) The most common measures of number sense are confounded by other cognitive demands, which drive key correlations. (4) Number sense and symbolic number systems (Arabic digits, number words, and so on) rely on distinct neural mechanisms and follow independent developmental trajectories. The review follows each assertion with comments on future directions that may bring resolution to these issues.

Mathematical Symbols, Learning Experiences and Number Sense: A Study of Three Grades Primary School Students

This study investigated the characteristics of three grades primary school students on number sensing. The performances on number sensing are dissimilar with different learning materials. While learning experiences also contributed to their number sensing process. 40 students selected from three grades were required to finish selection tasks. All 40 students had to select numerical value larger one from two spatial cubes, or from two Arabic digits, or from two lowercase Chinese character numbers. The results indicate that all students had easier selection process with Arabic numbers, while they had relative greater difficulty in making selection on numerical value with spatial cubes. The comparisons on mean scores and on reaction times from different-grade students indicate that all fourth-grade students had more learning experiences to develop number sense well, therefore their performances were better than other two grades’.

Does writing handedness affect neural representation of symbolic number? An fMRI adaptation study

A key question in the field of numerical cognition is how the human brain represents numerical symbols (e.g., Arabic digits). A large body of research has implicated left parietal regions in symbolic number processing. One possible explanation for this lateralization of neural activity is the handedness of participants. Specifically, participants in neuroimaging research are almost exclusively right-handed. The current study sought to probe whether number representation in the brain is associated with hand preference for handwriting. To address this question, we used functional Magnetic Resonance Imaging (fMRI) to compare brain activation of a group of right-handed participants with a group of left-handed participants during the passive viewing of symbolic numbers. At the whole-brain level, the right-handers demonstrated a previously obtained left-lateralized effect within the intraparietal sulcus (IPS). The left-handers showed some evidence of reverse lateralization of this effect in the IPS. However, when the groups were statistically compared, we found no regions that demonstrated group-level differences. In a follow-up region of interest (ROI) analysis within the left and right parietal lobes, we calculated laterality indices for each participant. Results did not provide support for group differences in laterality within the right or left IPS ROI. These findings do not support the hypothesis that handwriting plays a role in the lateralization of symbolic number processing in the brain. Further research is needed to better understand the factors that lead to left-lateralization for symbolic number processing.

Contrasting left/right codes for response selection must not be necessarily associated with contrasting numerical features to get the SNARC

The SNARC effect consists of faster reaction times to small numerical magnitudes when manual responses are delivered in the left-side of space and to large magnitudes when responses are delivered in the right-side. This spatial compatibility effect points at the interaction between the representations of space and that of numbers. Several studies have highlighted that an important determinant for the production of the SNARC is the use of contrasting left/right spatial codes in the selection of motor responses. In these studies, one spatial code for response selection, e.g. “left”, is usually associated with one number feature, e.g. “lower than 5”, while the contrasting spatial code, e.g. “right”, is associated with the contrasting number feature, e.g. “higher than 5”. Using a task with intermixed number and letter targets, here we show that significant and reliable SNARC effects are also produced when: a) one spatial response is associated with the intra-categorical discrimination of a number feature (i.e. magnitude or parity) and the contrasting response with the simple detection of letter targets; b) or when one spatial response is associated with the intra-categorical discrimination of the position of a letter in the alphabet (i.e. before or after “m”) and the contrasting spatial response with the simple detection of numerical targets (Experiments 1 and 2). In contrast, no reliable SNARC is found when no intra-categorical number or letter discrimination is required and contrasting left/right spatial response codes are simply associated with the discrimination between numbers and letters, e.g. “push left if the target is a number/push right if it is a letter” (Experiment 3). In a final control test (Experiment 4), we found no SNARC when the magnitude or parity classification of Arabic digits presented at central fixation was made unimanually with a left side or a right side-key and no left/right contrast was present in response selection. These results show that the use of contrasting left/right spatial response codes elicits reliable SNARC effects independently of their assignment to contrasting number features and confirm the important role played by the use of spatial codes in the genesis of the number-space interaction.

Arabic digits and spoken number words: Timing modulates the cross-modal numerical distance effect

Moving seamlessly between spoken number words and Arabic digits is common in everyday life. In this study, we systematically investigated the correspondence between auditory number words and visual Arabic digits in adults. Auditory number words and visual Arabic digits were presented concurrently or sequentially and participants had to indicate whether they described the same quantity. We manipulated the stimulus onset asynchronies (SOAs) between the two stimuli (Experiment 1: −500 ms to +500 ms; Experiment 2: −200 ms to +200 ms). In both experiments, we found a significant cross-modal distance effect. This effect was strongest for simultaneous stimulus presentation and decreased with increasing SOAs. Numerical distance emerged as the most consistent significant predictor overall, in particular for simultaneous presentation. However, physical similarity between the stimuli was often a significant predictor of response times in addition to numerical distance, and at longer SOAs, physical similarity between the stimuli was the only significant predictor. This shows that SOA modulates the extent to which participants access quantity representations. Our results thus support the idea that a semantic quantity representation of auditory and visual numerical symbols is activated when participants perform a concurrent matching task, while at longer SOAs participants are more likely to rely on physical similarity between the stimuli. We also investigated whether individual differences in the efficiency of the cross-modal processing were related to differences in mathematical performance. Our results are inconclusive about whether the efficiency of cross-format numerical correspondence is related to mathematical competence in adults.

Direct evidence for linguistic influences in two-digit number processing.

Language-specific differences in number words influence number processing even in nonverbal numerical tasks. For instance, the unit-decade compatibility effect in two-digit number magnitude comparison (compatible number pairs [42_57: 4 < 5 and 2 < 7] are responded to faster than incompatible pairs [47_62: 4 < 6 but 7 > 2]) was shown to be influenced by the inversion of number words (e.g., in German the number word for 42 is zweiundvierzig [literally: two-and-forty]). In two studies, we used articulatory suppression to investigate whether previously observed cross-linguistic differences in two-digit number processing are indeed driven by differences in number word formation. In a two-digit number comparison task, German- and English-speaking participants had to identify the larger of two numbers presented in Arabic digits. In Study 1, participants performed the same task twice, with and without articulatory suppression. In Study 2, the percentage of within-decade filler items (36_39) was manipulated additionally. As expected, in both studies between-groups differences in the compatibility effect disappeared under articulatory suppression irrespective of the percentage of fillers included. Furthermore, paralleling results of previous studies including 33% or less filler items, we found that the compatibility effect was larger in German compared with English speakers in the 20% filler condition. However, this pattern was reversed in the 50% filler condition in both studies. Thus, results provide first direct evidence for influences of verbal number word formation on symbolic number processing. Moreover, these new findings suggest that linguistic influences and those of cognitive control processes associated with characteristics of the stimulus set interact in symbolic number processing. (PsycINFO Database Record (c) 2019 APA, all rights reserved).