A New Multi-channels Sequence Recognition Framework Using Deep Convolutional Neural Network

Nowadays, a variety of sequences could be recorded and used with the rapid development of intelligent devices and sensors’ integrated technology. Several analysis of the sequences are based on the sequence recognition or classification and most of them are implemented via traditional machine learning models or their variants, such as Dynamic Time Warping, Hidden Markov Model and Support Vector Machine. Some of them could achieve a relatively high classification accuracy but with a time-consuming training process. Some other models are just the opposite. In this paper, we proposed a novel framework to solve the recognition task for sequences with multi-channels with a higher accuracy in less training time. In our framework, we designed a novel deep Convolutional Neural Network using “Data-Bands” as inputs. We conducted contrast experiments between our framework and several baseline methods and the results demonstrate that our framework could outperform state-of-art models