Skip to main content

A package for forensic face examination

Project description

forensicface

Install

pip install forensicface

Os arquivos onnx dos modelos de detecção (det_10g.onnx), pose (1k3d68.onnx) e gênero/idade (genderage.onnx) devem estar na pasta ~/.insightface/models/<model_name>/

O arquivo onnx do modelo de reconhecimento (adaface_ir101web12m.onnx) deve estar na pasta ~/.insightface/models/<model_name>/adaface/

O arquivo onnx do modelo de qualidade CR_FIQA (cr_fiqa_l.onnx) deve estar na pasta ~/.insightface/models/<model_name>/cr_fiqa/

O modelo padrão é denominado sepaelv2. Se você utilizar sepaelv2 como <model_name> não será preciso passar o nome do modelo durante a inicialização.

Como utilizar

Importação da classe ForensicFace:

from forensicface.app import ForensicFace

Instanciamento do ForensicFace:

ff = ForensicFace(det_size=320, use_gpu=True, extended=True)
Applied providers: ['CUDAExecutionProvider', 'CPUExecutionProvider'], with options: {'CPUExecutionProvider': {}, 'CUDAExecutionProvider': {'device_id': '0', 'gpu_mem_limit': '18446744073709551615', 'gpu_external_alloc': '0', 'gpu_external_free': '0', 'gpu_external_empty_cache': '0', 'cudnn_conv_algo_search': 'EXHAUSTIVE', 'cudnn_conv1d_pad_to_nc1d': '0', 'arena_extend_strategy': 'kNextPowerOfTwo', 'do_copy_in_default_stream': '1', 'enable_cuda_graph': '0', 'cudnn_conv_use_max_workspace': '1', 'tunable_op_enable': '0', 'enable_skip_layer_norm_strict_mode': '0', 'tunable_op_tuning_enable': '0'}}
find model: /home/rafael/.insightface/models/sepaelv2/1k3d68.onnx landmark_3d_68 ['None', 3, 192, 192] 0.0 1.0
Applied providers: ['CUDAExecutionProvider', 'CPUExecutionProvider'], with options: {'CPUExecutionProvider': {}, 'CUDAExecutionProvider': {'device_id': '0', 'gpu_mem_limit': '18446744073709551615', 'gpu_external_alloc': '0', 'gpu_external_free': '0', 'gpu_external_empty_cache': '0', 'cudnn_conv_algo_search': 'EXHAUSTIVE', 'cudnn_conv1d_pad_to_nc1d': '0', 'arena_extend_strategy': 'kNextPowerOfTwo', 'do_copy_in_default_stream': '1', 'enable_cuda_graph': '0', 'cudnn_conv_use_max_workspace': '1', 'tunable_op_enable': '0', 'enable_skip_layer_norm_strict_mode': '0', 'tunable_op_tuning_enable': '0'}}
find model: /home/rafael/.insightface/models/sepaelv2/det_10g.onnx detection [1, 3, '?', '?'] 127.5 128.0
Applied providers: ['CUDAExecutionProvider', 'CPUExecutionProvider'], with options: {'CPUExecutionProvider': {}, 'CUDAExecutionProvider': {'device_id': '0', 'gpu_mem_limit': '18446744073709551615', 'gpu_external_alloc': '0', 'gpu_external_free': '0', 'gpu_external_empty_cache': '0', 'cudnn_conv_algo_search': 'EXHAUSTIVE', 'cudnn_conv1d_pad_to_nc1d': '0', 'arena_extend_strategy': 'kNextPowerOfTwo', 'do_copy_in_default_stream': '1', 'enable_cuda_graph': '0', 'cudnn_conv_use_max_workspace': '1', 'tunable_op_enable': '0', 'enable_skip_layer_norm_strict_mode': '0', 'tunable_op_tuning_enable': '0'}}
find model: /home/rafael/.insightface/models/sepaelv2/genderage.onnx genderage ['None', 3, 96, 96] 0.0 1.0
set det-size: (320, 320)

Processamento básico de imagens

Obter pontos de referência, distância interpupilar, representação vetorial, a face alinhada com dimensão fixa (112x112), estimativas de sexo, idade, pose (pitch, yaw, roll) e qualidade.

results = ff.process_image_single_face("obama.png")
results.keys()
/home/rafael/miniconda3/envs/ffdev/lib/python3.10/site-packages/insightface/utils/transform.py:68: FutureWarning: `rcond` parameter will change to the default of machine precision times ``max(M, N)`` where M and N are the input matrix dimensions.
To use the future default and silence this warning we advise to pass `rcond=None`, to keep using the old, explicitly pass `rcond=-1`.
  P = np.linalg.lstsq(X_homo, Y)[0].T # Affine matrix. 3 x 4

dict_keys(['keypoints', 'ipd', 'embedding', 'norm', 'bbox', 'aligned_face', 'gender', 'age', 'pitch', 'yaw', 'roll', 'fiqa_score'])

Comparar duas imagens faciais e obter o escore de similaridade.

ff.compare("obama.png", "obama2.png")
0.8556093

Agregar embeddings de duas imagens faciais em uma única representação, com ponderação por qualidade

agg = ff.aggregate_from_images(["obama.png", "obama2.png"], quality_weight=True)
agg.shape
(512,)

Estimativa de qualidade CR-FIQA

Estimativa de qualidade pelo método CR-FIQA

Para desabilitar, instancie o forensicface com a opção extended = False:

ff = ForensicFace(extended=False)

Obs.: a opção extended = False também desabilita as estimativas de sexo, idade e pose.

good = ff.process_image("001_frontal.jpg")
bad = ff.process_image("001_cam1_1.jpg")
good["fiqa_score"], bad["fiqa_score"]
(2.3786173, 1.4386057)

Crédito dos modelos utilizados

  • Detecção, gênero (M/F), idade e pose (pitch, yaw, roll): insightface

  • Reconhecimento: adaface

  • Estimativa de qualidade: CR-FIQA

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

forensicface-0.3.7.tar.gz (16.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

forensicface-0.3.7-py3-none-any.whl (15.6 kB view details)

Uploaded Python 3

File details

Details for the file forensicface-0.3.7.tar.gz.

File metadata

  • Download URL: forensicface-0.3.7.tar.gz
  • Upload date:
  • Size: 16.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.10.11

File hashes

Hashes for forensicface-0.3.7.tar.gz
Algorithm Hash digest
SHA256 6b7cb1e24da4c5de01773fb5e6e01d7db808e63e185a36413e19247f586be100
MD5 db2c6e55eef41ecadf7a013f93d27364
BLAKE2b-256 564838875788b90e95edfa1f31b56c0961756b6e6b0745ca3ad2ce74027984ca

See more details on using hashes here.

File details

Details for the file forensicface-0.3.7-py3-none-any.whl.

File metadata

  • Download URL: forensicface-0.3.7-py3-none-any.whl
  • Upload date:
  • Size: 15.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.10.11

File hashes

Hashes for forensicface-0.3.7-py3-none-any.whl
Algorithm Hash digest
SHA256 62422d6e6447b50784385dcbc767e5ad4b716709d2fb80dc32f3c30a4c75d097
MD5 deb0b519f5737541e0cfdfaa84d49211
BLAKE2b-256 3e33127a5447762a0f51e896abcfe6561ceda69951a2e9cc33ac5fc6d6057b76

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page