A Recognition Method for Italian Alphabet Gestures Based on Convolutional Neural Network

Convolutional Neural Network(CNN) have achieved great success in image recognition and classification, but most of researches on gesture recognition are for English, there are very few identification studies for other small languages. An recognition method for Italian alphabet gestures based on CNN is proposed in this paper. First, an Italian alphabet gesture data set is created, including 22 letters, of which 4 letters (G, S, J, and Z) are excluded from the data set because of their dynamism, and a CNN model consisting of three blocks is established, convolutional and pooling layers are included in every block. Then weights and biases in the model are trained by the adaptive moment estimation algorithm called Adam to reduce the loss to the minimum step by step. Finally, we build a CNN model based on the proposed method on a deep learning platform named Keras, testing images are selected from the established static dataset randomly, the results of multiple experiments show that the recognition rate can reach up to 94%.

[1]  Vitaly Schetinin,et al.  Feature Extraction with GMDH-Type Neural Networks for EEG-Based Person Identification , 2017, Int. J. Neural Syst..

[2]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[3]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  Sunil Vadera,et al.  A convolutional neural network to classify American Sign Language fingerspelling from depth and colour images , 2017, Expert Syst. J. Knowl. Eng..

[5]  E. Kiran Kumar,et al.  Selfie Continuous Sign Language Recognition with Neural Network Classifier , 2018 .

[6]  Jun Sun,et al.  Invariant feature extraction for image classification via multi-channel convolutional neural network , 2017, 2017 International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS).

[7]  Adrian Akmajian,et al.  Sentence types and the form-function fit , 1984 .

[8]  Yong Hu,et al.  Sign Language Fingerspelling Recognition Using Depth Information and Deep Belief Networks , 2018, Int. J. Pattern Recognit. Artif. Intell..

[9]  Stefano Nolfi,et al.  Good teaching inputs do not correspond to desired responses in ecological neural networks , 1994, Neural Processing Letters.

[10]  Jie Huang,et al.  Video-based Sign Language Recognition without Temporal Segmentation , 2018, AAAI.

[11]  Li Zhang,et al.  Feature Extraction Based on Support Vector Data Description , 2018, Neural Processing Letters.

[12]  Hongxun Yao,et al.  Hierarchical semantic image matching using CNN feature pyramid , 2018, Comput. Vis. Image Underst..

[13]  William Chan,et al.  Transferring knowledge from a RNN to a DNN , 2015, INTERSPEECH.

[14]  Carlo Geraci,et al.  Visible degrees in Italian Sign Language , 2018 .

[15]  Md. Mohiminul Islam,et al.  Real time Hand Gesture Recognition using different algorithms based on American Sign Language , 2017, 2017 IEEE International Conference on Imaging, Vision & Pattern Recognition (icIVPR).

[16]  Andrzej Cichocki,et al.  Group Component Analysis for Multiblock Data: Common and Individual Feature Extraction , 2012, IEEE Transactions on Neural Networks and Learning Systems.