IEEE Sensors Journal · 2016 · 72 citations · 35 references
ComputingComputer ScienceArtificial IntelligenceComputer VisionMachine LearningMotion CapturePattern RecognitionText SegmentationWord RecognitionHandwritten Character RecognitionSegmentationStroke LengthWord Recognition AccuracyWord Segmentation
Recognition of 3‑D text drawn by fingers with Leap Motion is difficult because of jitter, non‑uniform character sizes, and continuous strokes that connect characters, words, and lines. The study investigates segmentation and recognition of text recorded using the Leap Motion sensor. Segmentation of continuous text into words is achieved by a heuristic based on stroke length between successive words, and each word is recognized by sequential classifiers—hidden Markov models and bidirectional LSTM neural networks—trained on a dataset of 560 Latin sentences from ten participants. Word segmentation achieved 78.2 % accuracy, while word recognition reached 86.88 % with BLSTM‑NN and 81.25 % with HMM.
Recognition of 3-D texts drawn by fingers using Leap motion sensor can be challenging for existing text recognition frameworks. The texts sensed by Leap motion device are different from traditional offline and on-line writing systems. This is because of frequent jitters and non-uniform character sizes while writing using Leap motion interface. Moreover, because of air writing, characters, words, and lines are usually connected by continuous stroke that makes it difficult to recognize. In this paper, we present a study of segmentation and recognition of text recorded using Leap motion sensor. The segmentation task of continuous text into words is performed using a heuristic analysis of stroke length between two successive words. Next, the recognition of each segmented word is performed using sequential classifiers. In this paper, we have performed 3-D text recognition using hidden Markov model (HMM) and bidirectional long short-term memory neural networks (BLSTM-NNs). We have created a data set consisting of 560 Latin sentences drawn by ten participants using Leap motion sensor for experiments. An accuracy of 78.2% has been obtained in word segmentation, whereas 86.88% and 81.25% accuracies have been recorded in word recognition using BLSTM-NN and HMM classifiers, respectively.
35