HFNet - 3.5 English

Vitis AI Library User Guide (UG1354)

Document ID
Release Date
3.5 English

HF-Net is a hierarchical localization approach based on a monolithic CNN that simultaneously predicts local features and global descriptors for accurate 6-DoF localization. 6-DoF visual localization method is accurate, scalable, and efficient, using HF-Net, a monolithic deep neural network for descriptor extraction. The proposed solution achieves state-of-the-art accuracy on several large-scale public benchmarks while running in real-time. For more details about HF-Net, refer to https://arxiv.org/abs/1812.03506.

The following image shows the result of HFNet.

Figure 1. HFNet Example

The following table lists the HFNet models supported by the Vitis AI Library.

Table 1. HFNet Models
No Model Name Framework
1 hfnet_tf TensorFlow