New Technique withConvolution Neural Networks (R- CNN's) Model for Hand Detection
Raad Ahmed Mohamed, Karim Q. Hussein · Turkish Journal of Computer and Mathematics Education (TURCOMAT) · 2021
The ability to listen and say the word are the most important aspects ofcommunication, but many of us are unlucky because we were not born with this skill from God.These people are deaf and dumb. Many studies are currently being conducted to address thedifficulties that these members of our society encounter in communicating with ordinary people.It is extremely difficult for mute (deaf and dumb) people to communicate their information to thegeneral public. Because the average person is not adequately prepared to grasp various sign. Itgets very difficult to communicate between these two types of people. Our research is only forthe purpose of assisting these mute (deaf and dumb) people in leading a better life. Manycomputer vision tasks involving human hands, such as hand pose estimation, hand identificationof gestures, human behavior analysis, and so on, are performed by humans. include handdetection as a critical pre-processing technique. However, due to the diverse appearancediversities of human hands, such Strong diffraction, weak resolution, fluctuating levels of light, avariety of hand gestures, and complicated interactions between hands and things or other handsare all factors to consider. such as (different hand forms, hand tracking, skin colors, scales,illuminations, orientations, gesture analgesia), accurately detecting hands is a difficult activity. incolor pictures, as well as Interaction between humans and machines, recognizing of signlanguages and so on). To address this problem, a region-based convolution neural networks (RCNN's)was used, in which hand regions are discovered and hand appearances are recreatedsimultaneously using attributes derived from a region proposal. The R-CNN was shown to besuitable for hand gesture detection with acceptable error.