Update README.md
Browse files
README.md
CHANGED
@@ -2,7 +2,9 @@
|
|
2 |
license: mit
|
3 |
---
|
4 |
|
5 |
-
|
|
|
|
|
6 |
|
7 |
Title: *Body-Part Joint Detection and Association via Extended Object Representation*
|
8 |
|
@@ -10,17 +12,40 @@ Paper link: https://arxiv.org/abs/2212.07652
|
|
10 |
|
11 |
Code link: https://github.com/hnuzhy/BPJDet
|
12 |
|
13 |
-
|
14 |
* BPJDet-S --> [ch_face_s_1536_e150_best_mMR.pt](./ch_face_s_1536_e150_best_mMR.pt)
|
15 |
* BPJDet-M --> [ch_face_m_1536_e150_best_mMR.pt](./ch_face_m_1536_e150_best_mMR.pt)
|
16 |
* BPJDet-L --> [ch_face_l_1536_e150_best_mMR.pt](./ch_face_l_1536_e150_best_mMR.pt)
|
17 |
|
18 |
-
|
19 |
* BPJDet-S --> [bh_hand_s_1536_e100_best.pt](./bh_hand_s_1536_e100_best.pt)
|
20 |
* BPJDet-M --> [bh_hand_m_1536_e100_best.pt](./bh_hand_m_1536_e100_best.pt)
|
21 |
* BPJDet-L --> [bh_hand_l_1536_e100_best.pt](./bh_hand_l_1536_e100_best.pt)
|
22 |
|
23 |
-
|
24 |
* BPJDet-S --> [ch_head_s_1536_e150_best_mMR.pt](./ch_head_s_1536_e150_best_mMR.pt)
|
25 |
* BPJDet-M --> [ch_head_m_1536_e150_best_mMR.pt](./ch_head_m_1536_e150_best_mMR.pt)
|
26 |
* BPJDet-L --> [ch_head_l_1536_e150_best_mMR.pt](./ch_head_l_1536_e150_best_mMR.pt)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
license: mit
|
3 |
---
|
4 |
|
5 |
+
# BPJDet
|
6 |
+
|
7 |
+
Pretrained models of our method **BPJDet**
|
8 |
|
9 |
Title: *Body-Part Joint Detection and Association via Extended Object Representation*
|
10 |
|
|
|
12 |
|
13 |
Code link: https://github.com/hnuzhy/BPJDet
|
14 |
|
15 |
+
## Body-Face Task (trained on CrowdHuman)
|
16 |
* BPJDet-S --> [ch_face_s_1536_e150_best_mMR.pt](./ch_face_s_1536_e150_best_mMR.pt)
|
17 |
* BPJDet-M --> [ch_face_m_1536_e150_best_mMR.pt](./ch_face_m_1536_e150_best_mMR.pt)
|
18 |
* BPJDet-L --> [ch_face_l_1536_e150_best_mMR.pt](./ch_face_l_1536_e150_best_mMR.pt)
|
19 |
|
20 |
+
## Body-Hand Task (trained on BodyHands)
|
21 |
* BPJDet-S --> [bh_hand_s_1536_e100_best.pt](./bh_hand_s_1536_e100_best.pt)
|
22 |
* BPJDet-M --> [bh_hand_m_1536_e100_best.pt](./bh_hand_m_1536_e100_best.pt)
|
23 |
* BPJDet-L --> [bh_hand_l_1536_e100_best.pt](./bh_hand_l_1536_e100_best.pt)
|
24 |
|
25 |
+
## Body-Head Task (trained on CrowdHuman)
|
26 |
* BPJDet-S --> [ch_head_s_1536_e150_best_mMR.pt](./ch_head_s_1536_e150_best_mMR.pt)
|
27 |
* BPJDet-M --> [ch_head_m_1536_e150_best_mMR.pt](./ch_head_m_1536_e150_best_mMR.pt)
|
28 |
* BPJDet-L --> [ch_head_l_1536_e150_best_mMR.pt](./ch_head_l_1536_e150_best_mMR.pt)
|
29 |
+
|
30 |
+
|
31 |
+
***
|
32 |
+
|
33 |
+
# BPJDetPlus
|
34 |
+
|
35 |
+
Pretrained models of our method **BPJDetPlus** (the extended journal version of BPJDet)
|
36 |
+
|
37 |
+
Title: *BPJDet: Extended Object Representation for Generic Body-Part Joint Detection*
|
38 |
+
|
39 |
+
Paper link: https://arxiv.org/abs/2304.10765
|
40 |
+
|
41 |
+
Code link: https://github.com/hnuzhy/BPJDet/tree/BPJDetPlus
|
42 |
+
|
43 |
+
## Body-Parts Task (trained on COCOHumanParts)
|
44 |
+
* BPJDetPlus-S --> [hp_parts_s_1536_e150_best_mMR.pt](./hp_parts_s_1536_e150_best_mMR.pt)
|
45 |
+
* BPJDetPlus-M --> [hp_parts_m_1536_e150_best_mMR.pt](./hp_parts_m_1536_e150_best_mMR.pt)
|
46 |
+
* BPJDetPlus-L --> [hp_parts_l_1536_e150_best_mMR.pt](./hp_parts_l_1536_e150_best_mMR.pt)
|
47 |
+
|
48 |
+
## Body-Parts Task (trained on COCOHumanParts)
|
49 |
+
* BPJDetPlus-S --> [ch_contact_s_1536_e100_best.pt](./ch_contact_s_1536_e100_best.pt)
|
50 |
+
* BPJDetPlus-M --> [ch_contact_m_1536_e100_best.pt](./ch_contact_m_1536_e100_best.pt)
|
51 |
+
* BPJDetPlus-L --> [ch_contact_l_1536_e100_best.pt](./ch_contact_l_1536_e100_best.pt)
|