Procedia Computer Science · 2019 · 26 citations · 9 references
ComputingComputer ScienceComputer VisionMachine VisionNatural Language ProcessingEdge DetectionFeature ExtractionGesture RecognitionLetter RecognitionPattern RecognitionShape AnalysisAccuracyHandwriting RecognitionImage Feature ExtractionSegmentationShape DetectionShape Feature ExtractionShape FeaturesSign Language Recognition
In this research the proposes a method of recognition of BISINDO letters based on hand-shape features that hint every shape of BISINDO Letters. In outline, this method is divided into two parts: the first is part of formation database shape features of BISINDO letters A-Z and the second is part of BISINDO letters recognition. In the first section consist of hand-shape image acquisition that hint every BISINDO letters, segmentation process, edge detection process, feature extraction process that is probability value of hand-shape chain code occurrence and process of database feature formation. In the second section is consist of hand-shape image acquisition process as BISINDO letters query followed by segmentation process, edge detection process, hand-shape feature extraction and recognition process by using calculation difference in distance between query shape feature to each shape feature in database feature. The image acquisition process in two parts above conducted directly (real time) via Webcam connected to the computer device. The method above has been implemented into prototype of Bisindo letters recognition software interface. The experiment results show the accuracy level of BISINDO letter recognition (26 BISINDO letters A to Z) which is reaching above 95%.
9
American sign language recognition with the kinect
Zahoor Zafrulla, Helene Brashear, Thad Starner et al. · 2011 · 373 citations
Comparative Study of Skin Color Detection and Segmentation in HSV and YCbCr Color Space
Khamar Basha Shaik, P. Ganesan, V. Kalist et al. · Procedia Computer Science · 2015 · 347 citations · Full text