Hugging Face's logo

STMicroelectronics
/
CNN2D-ST-HandPosture-model

Model card Files
xet
Community
FBAGSTM commited on
Commit
9b13645
·
verified ·
1 Parent(s): 6fe3f6e

Release AI-ModelZoo-4.0.0

Browse files
Files changed (1) hide show
  1. README.md +12 -14
README.md CHANGED
@@ -5,7 +5,7 @@ license_link: >-
5
  https://github.com/STMicroelectronics/stm32ai-modelzoo/raw/refs/heads/main/hand_posture/LICENSE.md
6
 
7
  ---
8
- # CNN2D_ST_HandPosture model
9
 
10
  ## **Use case** : `Hand posture recognition`
11
 
@@ -56,36 +56,34 @@ For an Time of Flight frame resolution of 8x8 and P classes
56
 
57
  ## Metrics
58
 
59
- Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
60
 
61
 
62
  ### Reference memory footprint based on ST_VL53LxCX_handposture_dataset (see Accuracy for details on dataset)
63
 
64
- | Model | Format | Input Shape | Series | Activation RAM (KiB) | Runtime RAM (KiB) | Weights Flash (KiB) | Code Flash (KiB) | Total RAM (KiB) | Total Flash (KiB) | STM32Cube.AI version |
65
- |:-----------------:|:------:|:-----------:|:-------:|:--------------:|:-----------:|:-------------:|:----------:|:-----------:|:-----------:|:---------------------:|
66
- | [CNN2D_ST_HandPosture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/CNN2D_ST_HandPosture/ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/CNN2D_ST_HandPosture_8classes/CNN2D_ST_HandPosture_8classes.h5) | FLOAT32 | 8 x 8 x 2 | STM32F4 | 1.07 | 2.08 | 10.75 | 14.37 | 3.15 | 25.12 | 10.2.0 |
67
- | [CNN2D_ST_HandPosture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/CNN2D_ST_HandPosture/ST_pretrainedmodel_custom_dataset/ST_VL53L5CX_handposture_dataset/CNN2D_ST_HandPosture_8classes/CNN2D_ST_HandPosture_8classes.h5) | FLOAT32 | 8 x 8 x 2 | STM32F4 | 1.07 | 2.08 | 10.75 | 14.37 | 3.15 | 25.12 | 10.2.0 |
68
 
69
 
70
  ### Reference inference time based on ST_VL53LxCX_handposture_dataset (see Accuracy for details on dataset)
71
 
72
 
73
- | Model | Format | Resolution | Board | Frequency | Inference time (ms) | STM32Cube.AI version |
74
  |:-----------------:|:------:|:----------:|:----------------:|:-------------:|:-------------------:|:---------------------:|
75
- | [CNN2D_ST_HandPosture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/CNN2D_ST_HandPosture/ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/CNN2D_ST_HandPosture_8classes/CNN2D_ST_HandPosture_8classes.h5) | FLOAT32 | 8 x 8 x 2 | STM32F401 | 84 MHz | 1.54 ms | 10.2.0 |
76
- | [CNN2D_ST_HandPosture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/CNN2D_ST_HandPosture/ST_pretrainedmodel_custom_dataset/ST_VL53L5CX_handposture_dataset/CNN2D_ST_HandPosture_8classes/CNN2D_ST_HandPosture_8classes.h5) | FLOAT32 | 8 x 8 x 2 | STM32F401 | 84 MHz | 1.53 ms | 10.2.0 |
77
 
78
  ### Accuracy with ST_VL53LxCX_handposture_dataset
79
 
80
  Number of classes: 8 [None, FlatHand, Like, Dislike, Fist, Love, BreakTime, CrossHands]. Training dataset number of frames: 3,031. Test dataset number of frames: 1146.
81
 
82
 
83
- | Model | Format | Resolution | Accuracy |
84
- |:-----------------:|:------:|:----------:|:----------------:|
85
- | [CNN2D_ST_HandPosture](ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/CNN2D_ST_HandPosture_8classes/CNN2D_ST_HandPosture_8classes.h5) | FLOAT32 | 8 x 8 x 2 | 99.43 % |
86
- | [CNN2D_ST_HandPosture](ST_pretrainedmodel_custom_dataset/ST_VL53L5CX_handposture_dataset/CNN2D_ST_HandPosture_8classes/CNN2D_ST_HandPosture_8classes.h5) | FLOAT32 | 8 x 8 x 2 | 97.17 % |
87
 
88
 
89
  ## Retraining and Integration in a simple example:
90
 
91
- Please refer to the stm32ai-modelzoo-services GitHub [here](https://github.com/STMicroelectronics/stm32ai-modelzoo-services)
 
5
  https://github.com/STMicroelectronics/stm32ai-modelzoo/raw/refs/heads/main/hand_posture/LICENSE.md
6
 
7
  ---
8
+ # st_cnn2d_handposture model
9
 
10
  ## **Use case** : `Hand posture recognition`
11
 
 
56
 
57
  ## Metrics
58
 
59
+ Measures are done with default STEdge AI Dev Cloud configuration with enabled input / output allocated option.
60
 
61
 
62
  ### Reference memory footprint based on ST_VL53LxCX_handposture_dataset (see Accuracy for details on dataset)
63
 
64
+ | Model | Format | Input Shape | Series | Activation RAM (KiB) | Runtime RAM (KiB) | Weights Flash (KiB) | Code Flash (KiB) | Total RAM (KiB) | Total Flash (KiB) | STEdge AI Core version |
65
+ |:--------------------------------------------------------------------------------------------------------------------------------------------------------:|:------:|:-----------:|:-------:|:--------------:|:-----------:|:-------------:|:----------:|:-----------:|:-----------:|:---------------------:|
66
+ | [st_cnn2d_handposture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/st_cnn2d_handposture/ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/st_cnn2d_handposture_8classes/st_cnn2d_handposture_8classes.keras) | FLOAT32 | 8 x 8 x 2 | STM32F4 | 1.63 | 0.28 | 10.75 | 6.16 | 1.91 | 16.19 | 3.0.0 |
 
67
 
68
 
69
  ### Reference inference time based on ST_VL53LxCX_handposture_dataset (see Accuracy for details on dataset)
70
 
71
 
72
+ | Model | Format | Resolution | Board | Frequency | Inference time (ms) | STEdge AI Core version |
73
  |:-----------------:|:------:|:----------:|:----------------:|:-------------:|:-------------------:|:---------------------:|
74
+ | [st_cnn2d_handposture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/st_cnn2d_handposture/ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/st_cnn2d_handposture_8classes/st_cnn2d_handposture_8classes.keras) | FLOAT32 | 8 x 8 x 2 | STM32F401 | 84 MHz | 1.46 ms | 3.0.0 |
 
75
 
76
  ### Accuracy with ST_VL53LxCX_handposture_dataset
77
 
78
  Number of classes: 8 [None, FlatHand, Like, Dislike, Fist, Love, BreakTime, CrossHands]. Training dataset number of frames: 3,031. Test dataset number of frames: 1146.
79
 
80
 
81
+ | Model | Dataset |Format | Resolution | Accuracy (%) |
82
+ |:--------------------------------------------------------------------------------------------------------------------------------------------------------:|:-------------------------------:|:---------:|:----------:|:------------:|
83
+ | [st_cnn2d_handposture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/st_cnn2d_handposture/ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/st_cnn2d_handposture_8classes/st_cnn2d_handposture_8classes.keras) | ST_VL53L8CX_handposture_dataset | FLOAT32 | 8 x 8 x 2 | 98.47 |
84
+ | [st_cnn2d_handposture](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/hand_posture/st_cnn2d_handposture/ST_pretrainedmodel_custom_dataset/ST_VL53L5CX_handposture_dataset/st_cnn2d_handposture_8classes/st_cnn2d_handposture_8classes.keras) | ST_VL53L5CX_handposture_dataset | FLOAT32 | 8 x 8 x 2 | 99.21 |
85
 
86
 
87
  ## Retraining and Integration in a simple example:
88
 
89
+ Please refer to the stm32ai-modelzoo-services GitHub [here](https://github.com/STMicroelectronics/stm32ai-modelzoo-services)