I7 7700

Medical Image Computing and Computer Assisted Intervention 2016 dataset was used for training on the network catenary. 1542 images of the training set and 900 images of the test set were randomly selected from the dataset. The size of the input images was 521 pixels ∗ 512∗. The research was carried out in Ubuntu system, and the constructed FCN and deconvolution network were run on GPU (Intel I7-7700) with 16 GB memory and NVIDIA GTX1070 GPU configuration. The neural network used in the experiment converged after continuous forward propagation and backpropagation iterative learning, and the parameters were saved. In the test stage, only the test image was input, and the segmentation result was obtained through the discrimination and classification of each pixel by softmax classification layer (Figure 3). In this study, different parameter values were selected to test the sensitivity, and the accuracy of different parameters on the whole image segmentation was analyzed.

The VR system was set up as previously described (Huang et al., 2022 (link)). Specifically, the VR system used in this study (Shanghai VR-Sens Intelligent Technology Co. Ltd., Shanghai, China) consisted of a VR interface, a VR headset, two controllers, and two cameras (Fig. 2). The VR system was linked to a conventional gaming computer (CPU: Intel i7-7700 and GPU: NVIDIA GTX1060). The VR interface offered scenarios in virtual reality. Two tracking cameras were used to precisely determine the location of the VR headset. The two controllers were employed to interact with VR objects. As a data visualization tool, custom software (VR-SENS VR Implant Tutorial Software) was utilized. The VR-compatible files were previously generated computationally.