Recognising sign language

Machine learning

Abstract— To overcome the difficulty of designing a decent neural network architecture, this Process uses a convolution neural network for gesture detection and names the network Auto GesNet. To be more explicit, we fuse and pre-process three gesture recognition data sets first. Then we develop AutoGesNet's general architecture and search space. In addition, we apply reinforcement learning and and apply teaching techniques toautomatically create AutoGesNet's comprehensive design. Finally, the searching neural network is fine-tuned and retrained for two distinct input sizes. Experiments demonstrate that the retrained model is accurate. on the NUS Hand Posture Dataset II and our data collection. A network that performs well in relation to recognition accuracy. We will compare and merge Autogenetic in future development.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.