data preprocessing

[xz0305]

Hello, I would like to ask how you get the semantic map and depth map, is it BisenetV2?

[conallwang]

Hi, as mentioned in Sec. 6 of our paper

Based on their processed images, we further obtain facial parsing results for each image using an open-source algorithm [17] and a depth map for each frame using Metashape software [25].

Actually, The semantic map (i.e., the facial parsing results) is obtained using MaskGAN based on this repo, and the depth map is obtained based on the reconstruction algorithm of the Metashape software which you can download from here.

[xz0305]

This solved my problem, thank you very much

---

发件人: conallwang @.> 发送时间: 2024年10月28日 11:26 收件人: conallwang/MeGA @.> 抄送: xz0305 @.>; Author @.> 主题: Re: [conallwang/MeGA] data preprocessing (Issue #3)

Hi, as mentioned in Sec. 6 of our paper

Based on their processed images, we further obtain facial parsing results for each image using an open-source algorithm [17] and a depth map for each frame using Metashape software [25].

Actually, The semantic map (i.e., the facial parsing results) is obtained using MaskGAN based on this repohttps://github.com/switchablenorms/CelebAMask-HQ, and the depth map is obtained based on the reconstruction algorithm of the Metashape software which you can download from herehttps://www.agisoft.com/downloads/installer/.

― Reply to this email directly, view it on GitHubhttps://github.com/conallwang/MeGA/issues/3#issuecomment-2440469902, or unsubscribehttps://github.com/notifications/unsubscribe-auth/BB2UWHL22L2ILK2GJF6UK2TZ5WVHPAVCNFSM6AAAAABQWPFHCSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINBQGQ3DSOJQGI. You are receiving this because you authored the thread.Message ID: @.***>