README.md

Dataset Description:

X-Humans Dataset contains high-quality textured human scans, rendered RGB-D images, and SMPL[-X] registrations. Totally, it has 20 subjects, 233 sequences, 35,427 frames. It includes individuals with a variety of body postures, hand gestures, and facial expressions, as well as diverse clothing styles, hairstyles, genders, and ages.

Name/ID	Gender	#Seq (Train)	#Seq (Test)	#Frames (Train)	#Frames (Test)
00016	male	9	2	1330	296
00017	male	9	2	1339	299
00018	male	9	2	1340	300
00019	female	7	2	1015	273
00020	male	10	3	1721	376
00021	female	6	1	1000	150
00024	male	12	3	1799	394
00025	female	9	2	1335	300
00027	female	10	2	1439	300
00028	male	10	3	1444	448
00034	male	10	2	1411	300
00035	male	10	2	1322	300
00036	female	10	3	1464	449
00039	female	10	2	1384	292
00041	male	12	2	1727	300
00058	female	10	2	1426	295
00085	female	8	2	1595	330
00086	female	10	2	1738	360
00087	male	9	2	1780	360
00088	male	10	2	1396	300
Total	/	190	43	29005	6422

File structure:

X_Humans
└── subject ID/
    ├── gender.txt: gender
    ├── mean_shape_smpl.npy: SMPL 'betas' (10,)
    ├── mean_shape_smplx.npy: SMPL-X 'betas' (10,)
    └── mode(train/test)/
        └── Sequence ID/
            ├── meshes_pkl/
            │   ├── atlas-fxxxxx.pkl: low-res textures as pickle files (1024, 1024, 3)
            │   └── mesh-fxxxxx.pkl: 'vertices', 'normals', 'uvs', 'faces'
            ├── render/
            │   ├── depth/
            │   │   └── depth_xxxxxx.tiff, depth image
            │   ├── image/
            │   │   └── color_xxxxxx.png, RGB image
            │   └── cameras.npz: intrinsic and extrinsic of the virtual camera
            ├── SMPL/
            │   ├── mesh-fxxxxx_smpl.pkl: SMPL params 'global_orient' (3,), 'transl' (3,), 'body_pose' (69,), 'betas' (10,) (use mean_shape_smpl.npy instead)
            │   └── mesh-fxxxxx_smpl.ply: SMPL meshes
            └── SMPLX/
                ├── mesh-fxxxxx_smplx.pkl: SMPL-X params 'global_orient' (3,), 'transl' (3,), 'body_pose' (63,), 'left_hand_pose' (45,), 'right_hand_pose' (45,), 
                │                         'jaw_pose' (3,), 'leye_pose' (3,), 'reye_pose' (3,), 'expression' (10,), 'betas' (10,) (use mean_shape_smplx.npy instead)
                └── mesh-fxxxxx_smplx.ply: SMPL-X meshes

Code instruction:

• preprocess_XHumans.py: prepare X-Humans dataset for training X-Avatar https://github.com/Skype-line/X-Avatar
• vis_meshes.py: visualize X-Humans 3D scans and SMPL[-X] meshes with aitviewer https://github.com/eth-ait/aitviewer