person-attributes-recognition-crossroad-0234¶
Use Case and High-Level Description¶
This model presents a person attributes classification algorithm analysis scenario. The model consists of the ResNet-50 backbone and a head. For an input image with a pedestrian the model returns 7 values that are probabilities of the corresponding 7 attributes.
Specification¶
Metric |
Value |
---|---|
Pedestrian pose |
Standing person |
Occlusion coverage |
<20% |
Min object width |
80 pixels |
Supported attributes |
|
GFlops |
2.167 |
MParams |
23.510 |
Source framework |
PyTorch* |
Accuracy¶
Attribute |
F1 |
---|---|
|
0.92 |
|
0.44 |
|
0.74 |
|
0.45 |
|
0.89 |
|
0.84 |
|
NA |
Inputs¶
Image, name: input
, shape: 1, 3, 160, 80
in the format 1, C, H, W
, where:
C
- number of channelsH
- image heightW
- image width
The expected color order is BGR
.
Outputs¶
The net output is a blob named attributes
with shape 1, 7
across seven attributes: [is_male
, has_bag
, has_hat
, has_longsleeves
, has_longpants
, has_longhair
, has_coat_jacket
]. Value > 0.5 means that the corresponding attribute is present.
Demo usage¶
The model can be used in the following demos provided by the Open Model Zoo to show its capabilities:
Legal Information¶
[*] Other names and brands may be claimed as the property of others.