Figure 5 | Scientific Reports

Figure 5

From: Exploiting domain transformation and deep learning for hand gesture recognition using a low-cost dataglove

Figure 5

Proposed network architectures: (a) Overall diagram of the proposed architecture. For static gestures, the sensor channels are processed by parallel 1D ConvNet blocks. For dynamic gestures, the accelerations are first converted into spatial projection images and features are extracted from them using the pre-trained MobileNetV2 network, (b) the architecture of the 1D ConvNet Blocks, and (c) the architecture of MobileNetV2.

Back to article page