MishaF commited on
Commit
e2d41dc
1 Parent(s): 59722dd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +25 -14
README.md CHANGED
@@ -32,20 +32,31 @@ Bria-RMBG model was trained over 12000 high quality, high resolution, fully lice
32
  The training set as well as the validation benchmark if a holistic representation of the commercial world containing a distribution of general stock images, eComm, gaming and ads.
33
 
34
  Distribution of images:
35
- only people - 17.35%
36
- people with objects/animals - 25.24%
37
- only objects - 45.11%
38
- only animals - 1.89%
39
- only text - 2.52%
40
- people/objects/animals with text - 8.52%
41
-
42
- photorealistic - 87.70%
43
- non photorealistic - 12.30%
44
- solid bg - 47.95%
45
- non solid bg - 52.05%
46
- single main object in the foreground - 51.42%
47
- multiple objects in the foreground- 48.58%
48
-
 
 
 
 
 
 
 
 
 
 
 
49
 
50
  All images were manualy labeled pixel-wise accuratly.
51
 
 
32
  The training set as well as the validation benchmark if a holistic representation of the commercial world containing a distribution of general stock images, eComm, gaming and ads.
33
 
34
  Distribution of images:
35
+ | Category | Distribution |
36
+ | -----------------------------------| -----------------------------------:|
37
+ | Objects only | 45.11 |
38
+ | People with objects/animals | 25.24% |
39
+ | People only | 17.35% |
40
+ | people/objects/animals with text | 8.52% |
41
+ | Text only | 2.52% |
42
+ | Animals only | 1.89% |
43
+
44
+ | Category | Distribution |
45
+ | -----------------------------------| -----------------------------------------:|
46
+ | Photorealistic | 87.70% |
47
+ | Non-Photorealistic | 12.30% |
48
+
49
+
50
+ | Category | Distribution |
51
+ | -----------------------------------| -----------------------------------:|
52
+ | Non Solid Background | 52.05% |
53
+ | Solid Background | 47.95%
54
+
55
+
56
+ | Category | Distribution |
57
+ | -----------------------------------| -----------------------------------:|
58
+ | Single main foreground object | 51.42% |
59
+ | Multiple objects in the foreground | 48.58% |
60
 
61
  All images were manualy labeled pixel-wise accuratly.
62