☆ Export paths: ☆

export WEIGHTS="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/model-zoo/nobi_model_v3/scaled_nobi_pose_v3.weights"
export NAMES="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/model-zoo/nobi_model_v3/scaled_nobi_pose_v3.names"
export CFG="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/model-zoo/nobi_model_v3/scaled_nobi_pose_v3.cfg"
export SAVE_DIR="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/nobi-hw-videocapture/EMoi///"
export ENGINE="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/model-zoo/nobi_model_v3/scaled_nobi_pose_v3.engine"
export ALPHAPOSE_MODEL="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/AlphaPose/AlphaPose_TorchScript/model-zoo/fast_pose_res50/fast_res50_256x192.jit"
export TABULAR_MODEL="/mnt/2B59B0F32ED5FBD7/Projects/KIKAI/model-zoo/nobi_tabular_v1/tabular_v1.jit"

• If you using darknet for detector:

cd darknet
make -j$(nproc)

☆ Streamming ☆

BASH

Build Streamming 4 Camera Test

rm -rf .cmake/ *
cmake -DFOUR_CAMS=ON -DCAM_ID_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App

Build Streamming 4 Video Source Test

rm -rf .cmake/ *
cmake -DFOUR_CAMS=ON -DVIDEO_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App

Build Streaming From Camera Test

rm -rf .cmake/ *
cmake -DSINGLE_CAM=ON -DCAM_ID_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App

Build Streaming From Video Soure Test

rm -rf .cmake/ *
cmake -DSINGLE_CAM=ON -DVIDEO_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App

☆ Video Inference ☆

BASH

Build test object detection by TensorRT model on a sample video path like darknet

rm -rf .cmake/ *
cmake -DINFERENCE_VIDEO=ON -DVIDEO_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App

Build test object detection by darknet C++ model on a sample video path like darknet

rm -rf .cmake/ *
cmake -DINFERENCE_VIDEO=ON -DINFERENCE_DARKNET=ON -DVIDEO_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App

Build test object detection by TensorRT model on a sample video path like darknet, and support pose estimation

rm -rf .cmake/ *
cmake -DINFERENCE_VIDEO=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON ..
cmake --build . --config Release
./Nobi_App

Build test object detection by darknet C++ model on a sample video path like darknet, and support pose estimation

rm -rf .cmake/ *
cmake -DINFERENCE_VIDEO=ON -DINFERENCE_DARKNET=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON ..
cmake --build . --config Release
./Nobi_App

☆ API like darknet ☆

RUN WITHOUT JSON

Build test object detection by TensorRT model on a specific image path like darknet

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build test object detection by darknet C++ model on a specific image path like darknet

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_DARKNET=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --thresh 0.3 --dont-show

Build test object detection by TensorRT model on a specific image path like darknet, and support pose estimation

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_ALPHAPOSE_TORCH=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build test object detection by darknet C++ model on a specific image path like darknet, and support pose estimation

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_DARKNET=ON -DINFERENCE_ALPHAPOSE_TORCH=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --thresh 0.3 --dont-show

RUN WITH JSON

Build test object detection by TensorRT model on a sample video path like darknet and output Json

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build test object detection by darknet C++ model on a sample video path like darknet and output Json

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_DARKNET=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --thresh 0.3 --dont-show

Build test object detection by TensorRT model on a sample video path like darknet, and support pose estimation and output Json

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_ALPHAPOSE_TORCH=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build test object detection by TensorRT model on a sample video path like darknet, and support pose estimation, tabular learner and output Json

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_ALPHAPOSE_TORCH=ON -DINFERENCE_TABULAR_TORCH=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --tabular-jit ${TABULAR_MODEL} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build test object detection by darknet C++ model on a sample video path like darknet, and support pose estimation and output Json

rm -rf .cmake/ *
cmake -DTENSORRT_API=ON -DINFERENCE_DARKNET=ON -DINFERENCE_ALPHAPOSE_TORCH=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --thresh 0.3 --dont-show

☆ Nobi AI Camera API ☆

RUN WITHOUT JSON

Build Nobi API support object detection TensorRT

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DVIDEO_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --save-dir ${SAVE_DIR} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build Nobi API support object detection darknet C++

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DINFERENCE_DARKNET=ON -DVIDEO_EXAMPLES=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --save-dir ${SAVE_DIR} --thresh 0.3 --dont-show

Build Nobi API support object detection TensorRT and pose estimation

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --save-dir ${SAVE_DIR} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build Nobi API support object detection darknet C++ and pose estimation

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DINFERENCE_DARKNET=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --save-dir ${SAVE_DIR} --thresh 0.3 --dont-show

RUN WITH JSON

Build Nobi API support object detection TensorRT and output Json

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DVIDEO_EXAMPLES=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --save-dir ${SAVE_DIR} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build Nobi API support object detection darknet C++ and output Json

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DINFERENCE_DARKNET=ON -DVIDEO_EXAMPLES=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --save-dir ${SAVE_DIR} --thresh 0.3 --dont-show

Build Nobi API support object detection TensorRT and pose estimation and output Json

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --save-dir ${SAVE_DIR} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build Nobi API support object detection TensorRT and pose estimation, tabular learner and output Json

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON -DINFERENCE_TABULAR_TORCH=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --engine-file ${ENGINE} --label-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --tabular-jit ${TABULAR_MODEL} --save-dir ${SAVE_DIR} --dims 512 512 --obj-thres 0.3 --nms-thres 0.4 --type-yolo csp --dont-show

Build Nobi API support object detection darknet C++ and pose estimation and output Json

rm -rf .cmake/ *
cmake -DNOBI_CAMERA_AI_API=ON -DINFERENCE_DARKNET=ON -DVIDEO_EXAMPLES=ON -DINFERENCE_ALPHAPOSE_TORCH=ON -DJSON=ON ..
cmake --build . --config Release
./Nobi_App --weights-file ${WEIGHTS} --cfg-file ${CFG} --names-file ${NAMES} --alphapose-jit ${ALPHAPOSE_MODEL} --save-dir ${SAVE_DIR} --thresh 0.3 --dont-show