Machine Vision Inspection Systems, Machine Learning-Based Approaches. Группа авторов

25 4.00% 62.80% Tengwar 24 3.33% 62.08% ULOG 25 3.60% 61.60% Syriac (Serrto) 22 6.36% 58.64% Atlantean 25 2.80% 58.00% Avesta 25 5.20% 57.60% Cyrillic 44 2.05% 57.05% Sinhala 60 1.00% 56.22% Ge`ez 25 1.60% 52.40% Mongolian 29 4.83% 52.07% Glagolitic 44 1.82% 50.68% Manipuri 39 3.08% 50.51% Malayalam 46 3.26% 45.87% Tibetan 41 2.93% 45.61% Sylheti 27 4.07% 40.37% Gurmukhi 44 2.27% 38.41% Oriya 45 1.56% 33.33% Kannada 40 1.00% 29.25%

For further analysis, we consider the alphabet models with the same number of characters and those have shown the highest and lowest classification accuracies. Consider the characters of Gurmukhi (38.41% accuracy) and Cyrillic (57.05% accuracy), which has the same number of characters (44), but accuracy differs by 18.64%. The accuracy difference could be due to the structural similarity between characters in those alphabets. Figure 2.6 shows the two alphabets. Due to the same reason, we get lower accuracies for within language classification compared to mixed language n-way classification as described in Section 2.4.1.

      Further, in an attempt to boost the accuracies in classification, we have used n-shot learning, while keeping 10 images for each character in the alphabet as the reference set and 10 images for averaging the results. By this experiment, we obtained 7 to 15% accuracy improvement resulting in 94% highest accuracy for Aurek-Besh language and 40% lowest accuracy for Oriya language, respectively.

      2.4.3 MNIST Classification

The Omniglot dataset has more than 1,600-character classes, but has only 20 samples for each category. In contrast, MNIST dataset has 10 classes and 60,000 total training samples [30]. Since the proposed model of this study aims to learn an abstract knowledge about characters and extend it to identify new characters, by treating MNIST as a whole new alphabet with 10 characters, we could use the proposed capsule layers-based Siamese network model to apply classifications for MNIST dataset. Table 2.4 shows the accuracy values obtained by different MNIST models. Here, large neural networks have achieved more than 90% accuracy while the proposed capsule layers-based Siamese network model has given 76% accuracy with only 20 images.

Figure 2.6 Gurmukhi (left) and Cyrillic (right) alphabets.

Table 2.4 Accuracies of different MNIST models.

MNIST Model	Accuracy
1-Layer NN [18]	88%
2-layer NN [18]	95.3%
Large convolutional NN [25]	99.5%
Proposed capsule layer-based Siamese network (1-shot)	51%
Proposed capsule layer-based Siamese network (20-shot)	74.5%

      The MNIST dataset is a benchmark model for image classification algorithms and has been solved to get more than 90% accuracy as summarized in Table 2.4. These methods are based on deep neural networks and use all the 60K characters in the dataset.

Although the proposed capsule layers-based Siamese network model has shown only 51% accuracy with MNIST dataset, that has used only one sample for each digit class while other models have access to more than 60,000 samples. The proposed solution has improved this accuracy by using the same n-shot learning technique. By using 20 samples the accuracy is improved by 23.5% as depicted in Figure 2.7. Thus, the classification accuracy of MNIST dataset is improved from 51 to 74.5% by using a greater number of samples.

Figure 2.7 MNIST n-shot learning performance.

      2.4.4 Sinhala Language Classification

      One of the main goals in this research is evaluating the performance of one-shot learning for Sinhala language. Using deep learning approaches is not an option for Sinhala character recognition due to a lack of datasets. Sinhala language has 60 characters, making