EXCALORIO commited on
Commit
483e750
·
verified ·
1 Parent(s): beb2c4c

ficha: enlazar tambien los fijos que no van incrustados

Browse files
Files changed (1) hide show
  1. README.md +12 -0
README.md CHANGED
@@ -62,6 +62,8 @@ Full-resolution stills:
62
  ![pose_BabyishIdioticCockerspaniel_c0167](media/pose_BabyishIdioticCockerspaniel_c0167.png)
63
  ![pose_proudDungBeetleBoglenTerrier_c0239](media/pose_proudDungBeetleBoglenTerrier_c0239.png)
64
 
 
 
65
 
66
  - **2.26 px** torso error from running the pose at 384x288 instead of 1024, which is **7.8x** cheaper
67
  - **96.0 %** of delivered poses are measured; the rest are interpolated and **say so**
@@ -91,6 +93,8 @@ Full-resolution stills:
91
  ![seg_BabyishIdioticCockerspaniel_c0167](media/seg_BabyishIdioticCockerspaniel_c0167.png)
92
  ![seg_proudDungBeetleBoglenTerrier_c0239](media/seg_proudDungBeetleBoglenTerrier_c0239.png)
93
 
 
 
94
 
95
  - referee leakage into the mask measured **0.0000**
96
  - the mask is `(matting >= 0.5 OR seg != background) AND nearest-skeleton AND box guard`, and the loss is attributed to the three of them separately - a low recall does not say what to fix
@@ -116,6 +120,8 @@ Full-resolution stills:
116
  ![normal_BabyishIdioticCockerspaniel_c0167](media/normal_BabyishIdioticCockerspaniel_c0167.png)
117
  ![normal_proudDungBeetleBoglenTerrier_c0239](media/normal_proudDungBeetleBoglenTerrier_c0239.png)
118
 
 
 
119
 
120
  - outside the mask the value is **NaN and not 0**, because the zero vector is not a unit normal but it would average silently inside a convolution
121
 
@@ -139,6 +145,8 @@ Full-resolution stills:
139
  ![matting_BabyishIdioticCockerspaniel_c0167](media/matting_BabyishIdioticCockerspaniel_c0167.png)
140
  ![matting_proudDungBeetleBoglenTerrier_c0239](media/matting_proudDungBeetleBoglenTerrier_c0239.png)
141
 
 
 
142
 
143
  - the footer prints the **mean alpha per side**. A near-zero one is not a bug in the video: `matting` returns background when a second person is inside the crop, and then that fencer's silhouette is empty. A null has to say it is a null and why
144
 
@@ -162,6 +170,8 @@ Full-resolution stills:
162
  ![pointmap_BabyishIdioticCockerspaniel_c0167](media/pointmap_BabyishIdioticCockerspaniel_c0167.png)
163
  ![pointmap_proudDungBeetleBoglenTerrier_c0239](media/pointmap_proudDungBeetleBoglenTerrier_c0239.png)
164
 
 
 
165
 
166
  - it orders the two fencers in depth correctly in **76 %** of frames - and the **y coordinate of the free foot**, which costs nothing, gets it right **100 %** (28 of 28)
167
  - Sapiens 1 depth scored **73.2 %**, below the free baseline, on a newer model's predecessor
@@ -191,6 +201,8 @@ Full-resolution stills:
191
  ![foil_proudDungBeetleBoglenTerrier_c0163](media/foil_proudDungBeetleBoglenTerrier_c0163.png)
192
  ![foil_uptightSlothEveningBat_c0146](media/foil_uptightSlothEveningBat_c0146.png)
193
 
 
 
194
 
195
  - reach **1.10 m** wrist-to-tip. A foil blade is **0.90 m** on its own, and the first version used exactly that - which is wrong for a *threat* radius, because the threat starts at the wrist and the guard and the hand add the rest. It looked too short and it was
196
  - the axis is the **forearm**, `elbow -> hand root`. Using the hand axis instead put the cone more than 40 deg off in **63.6 %** of frames; the forearm brings that to **15.7 %**
 
62
  ![pose_BabyishIdioticCockerspaniel_c0167](media/pose_BabyishIdioticCockerspaniel_c0167.png)
63
  ![pose_proudDungBeetleBoglenTerrier_c0239](media/pose_proudDungBeetleBoglenTerrier_c0239.png)
64
 
65
+ More stills, same layer: [BabyishIdioticCockerspaniel_c0302](media/pose_BabyishIdioticCockerspaniel_c0302.png) | [BabyishIdioticCockerspaniel_c0315](media/pose_BabyishIdioticCockerspaniel_c0315.png) | [proudDungBeetleBoglenTerrier_c0096](media/pose_proudDungBeetleBoglenTerrier_c0096.png) | [proudDungBeetleBoglenTerrier_c0163](media/pose_proudDungBeetleBoglenTerrier_c0163.png) | [proudDungBeetleBoglenTerrier_c0230](media/pose_proudDungBeetleBoglenTerrier_c0230.png) | [proudDungBeetleBoglenTerrier_c0334](media/pose_proudDungBeetleBoglenTerrier_c0334.png) | [uptightSlothEveningBat_c0146](media/pose_uptightSlothEveningBat_c0146.png) | [uptightSlothEveningBat_c0187](media/pose_uptightSlothEveningBat_c0187.png) | [uptightSlothEveningBat_c0227](media/pose_uptightSlothEveningBat_c0227.png) | [uptightSlothEveningBat_c0303](media/pose_uptightSlothEveningBat_c0303.png) | [uptightSlothEveningBat_c0362](media/pose_uptightSlothEveningBat_c0362.png)
66
+
67
 
68
  - **2.26 px** torso error from running the pose at 384x288 instead of 1024, which is **7.8x** cheaper
69
  - **96.0 %** of delivered poses are measured; the rest are interpolated and **say so**
 
93
  ![seg_BabyishIdioticCockerspaniel_c0167](media/seg_BabyishIdioticCockerspaniel_c0167.png)
94
  ![seg_proudDungBeetleBoglenTerrier_c0239](media/seg_proudDungBeetleBoglenTerrier_c0239.png)
95
 
96
+ More stills, same layer: [BabyishIdioticCockerspaniel_c0302](media/seg_BabyishIdioticCockerspaniel_c0302.png) | [BabyishIdioticCockerspaniel_c0315](media/seg_BabyishIdioticCockerspaniel_c0315.png) | [proudDungBeetleBoglenTerrier_c0096](media/seg_proudDungBeetleBoglenTerrier_c0096.png) | [proudDungBeetleBoglenTerrier_c0163](media/seg_proudDungBeetleBoglenTerrier_c0163.png) | [proudDungBeetleBoglenTerrier_c0230](media/seg_proudDungBeetleBoglenTerrier_c0230.png) | [proudDungBeetleBoglenTerrier_c0334](media/seg_proudDungBeetleBoglenTerrier_c0334.png) | [uptightSlothEveningBat_c0146](media/seg_uptightSlothEveningBat_c0146.png) | [uptightSlothEveningBat_c0187](media/seg_uptightSlothEveningBat_c0187.png) | [uptightSlothEveningBat_c0227](media/seg_uptightSlothEveningBat_c0227.png) | [uptightSlothEveningBat_c0303](media/seg_uptightSlothEveningBat_c0303.png) | [uptightSlothEveningBat_c0362](media/seg_uptightSlothEveningBat_c0362.png)
97
+
98
 
99
  - referee leakage into the mask measured **0.0000**
100
  - the mask is `(matting >= 0.5 OR seg != background) AND nearest-skeleton AND box guard`, and the loss is attributed to the three of them separately - a low recall does not say what to fix
 
120
  ![normal_BabyishIdioticCockerspaniel_c0167](media/normal_BabyishIdioticCockerspaniel_c0167.png)
121
  ![normal_proudDungBeetleBoglenTerrier_c0239](media/normal_proudDungBeetleBoglenTerrier_c0239.png)
122
 
123
+ More stills, same layer: [BabyishIdioticCockerspaniel_c0302](media/normal_BabyishIdioticCockerspaniel_c0302.png) | [BabyishIdioticCockerspaniel_c0315](media/normal_BabyishIdioticCockerspaniel_c0315.png) | [proudDungBeetleBoglenTerrier_c0096](media/normal_proudDungBeetleBoglenTerrier_c0096.png) | [proudDungBeetleBoglenTerrier_c0163](media/normal_proudDungBeetleBoglenTerrier_c0163.png) | [proudDungBeetleBoglenTerrier_c0230](media/normal_proudDungBeetleBoglenTerrier_c0230.png) | [proudDungBeetleBoglenTerrier_c0334](media/normal_proudDungBeetleBoglenTerrier_c0334.png) | [uptightSlothEveningBat_c0146](media/normal_uptightSlothEveningBat_c0146.png) | [uptightSlothEveningBat_c0187](media/normal_uptightSlothEveningBat_c0187.png) | [uptightSlothEveningBat_c0227](media/normal_uptightSlothEveningBat_c0227.png) | [uptightSlothEveningBat_c0303](media/normal_uptightSlothEveningBat_c0303.png) | [uptightSlothEveningBat_c0362](media/normal_uptightSlothEveningBat_c0362.png)
124
+
125
 
126
  - outside the mask the value is **NaN and not 0**, because the zero vector is not a unit normal but it would average silently inside a convolution
127
 
 
145
  ![matting_BabyishIdioticCockerspaniel_c0167](media/matting_BabyishIdioticCockerspaniel_c0167.png)
146
  ![matting_proudDungBeetleBoglenTerrier_c0239](media/matting_proudDungBeetleBoglenTerrier_c0239.png)
147
 
148
+ More stills, same layer: [BabyishIdioticCockerspaniel_c0302](media/matting_BabyishIdioticCockerspaniel_c0302.png) | [BabyishIdioticCockerspaniel_c0315](media/matting_BabyishIdioticCockerspaniel_c0315.png) | [proudDungBeetleBoglenTerrier_c0096](media/matting_proudDungBeetleBoglenTerrier_c0096.png) | [proudDungBeetleBoglenTerrier_c0163](media/matting_proudDungBeetleBoglenTerrier_c0163.png) | [proudDungBeetleBoglenTerrier_c0230](media/matting_proudDungBeetleBoglenTerrier_c0230.png) | [proudDungBeetleBoglenTerrier_c0334](media/matting_proudDungBeetleBoglenTerrier_c0334.png) | [uptightSlothEveningBat_c0146](media/matting_uptightSlothEveningBat_c0146.png) | [uptightSlothEveningBat_c0187](media/matting_uptightSlothEveningBat_c0187.png) | [uptightSlothEveningBat_c0227](media/matting_uptightSlothEveningBat_c0227.png) | [uptightSlothEveningBat_c0303](media/matting_uptightSlothEveningBat_c0303.png) | [uptightSlothEveningBat_c0362](media/matting_uptightSlothEveningBat_c0362.png)
149
+
150
 
151
  - the footer prints the **mean alpha per side**. A near-zero one is not a bug in the video: `matting` returns background when a second person is inside the crop, and then that fencer's silhouette is empty. A null has to say it is a null and why
152
 
 
170
  ![pointmap_BabyishIdioticCockerspaniel_c0167](media/pointmap_BabyishIdioticCockerspaniel_c0167.png)
171
  ![pointmap_proudDungBeetleBoglenTerrier_c0239](media/pointmap_proudDungBeetleBoglenTerrier_c0239.png)
172
 
173
+ More stills, same layer: [BabyishIdioticCockerspaniel_c0302](media/pointmap_BabyishIdioticCockerspaniel_c0302.png) | [BabyishIdioticCockerspaniel_c0315](media/pointmap_BabyishIdioticCockerspaniel_c0315.png) | [proudDungBeetleBoglenTerrier_c0096](media/pointmap_proudDungBeetleBoglenTerrier_c0096.png) | [proudDungBeetleBoglenTerrier_c0163](media/pointmap_proudDungBeetleBoglenTerrier_c0163.png) | [proudDungBeetleBoglenTerrier_c0230](media/pointmap_proudDungBeetleBoglenTerrier_c0230.png) | [proudDungBeetleBoglenTerrier_c0334](media/pointmap_proudDungBeetleBoglenTerrier_c0334.png) | [uptightSlothEveningBat_c0146](media/pointmap_uptightSlothEveningBat_c0146.png) | [uptightSlothEveningBat_c0187](media/pointmap_uptightSlothEveningBat_c0187.png) | [uptightSlothEveningBat_c0227](media/pointmap_uptightSlothEveningBat_c0227.png) | [uptightSlothEveningBat_c0303](media/pointmap_uptightSlothEveningBat_c0303.png) | [uptightSlothEveningBat_c0362](media/pointmap_uptightSlothEveningBat_c0362.png)
174
+
175
 
176
  - it orders the two fencers in depth correctly in **76 %** of frames - and the **y coordinate of the free foot**, which costs nothing, gets it right **100 %** (28 of 28)
177
  - Sapiens 1 depth scored **73.2 %**, below the free baseline, on a newer model's predecessor
 
201
  ![foil_proudDungBeetleBoglenTerrier_c0163](media/foil_proudDungBeetleBoglenTerrier_c0163.png)
202
  ![foil_uptightSlothEveningBat_c0146](media/foil_uptightSlothEveningBat_c0146.png)
203
 
204
+ More stills, same layer: [BabyishIdioticCockerspaniel_c0302](media/foil_BabyishIdioticCockerspaniel_c0302.png) | [BabyishIdioticCockerspaniel_c0315](media/foil_BabyishIdioticCockerspaniel_c0315.png) | [proudDungBeetleBoglenTerrier_c0096](media/foil_proudDungBeetleBoglenTerrier_c0096.png) | [proudDungBeetleBoglenTerrier_c0230](media/foil_proudDungBeetleBoglenTerrier_c0230.png) | [proudDungBeetleBoglenTerrier_c0239](media/foil_proudDungBeetleBoglenTerrier_c0239.png) | [proudDungBeetleBoglenTerrier_c0334](media/foil_proudDungBeetleBoglenTerrier_c0334.png) | [uptightSlothEveningBat_c0187](media/foil_uptightSlothEveningBat_c0187.png) | [uptightSlothEveningBat_c0227](media/foil_uptightSlothEveningBat_c0227.png) | [uptightSlothEveningBat_c0303](media/foil_uptightSlothEveningBat_c0303.png) | [uptightSlothEveningBat_c0362](media/foil_uptightSlothEveningBat_c0362.png)
205
+
206
 
207
  - reach **1.10 m** wrist-to-tip. A foil blade is **0.90 m** on its own, and the first version used exactly that - which is wrong for a *threat* radius, because the threat starts at the wrist and the guard and the hand add the rest. It looked too short and it was
208
  - the axis is the **forearm**, `elbow -> hand root`. Using the hand axis instead put the cone more than 40 deg off in **63.6 %** of frames; the forearm brings that to **15.7 %**