tensorflowDetection2D module use tensorflow C++ API. The module detects using pre-trained models like VGG16, Inception, ResNet, MobileNet and other pre-trained models. The module build with cmake instead of bazel using tensorflow_cc. Also use YARP to send video source pre and post-procesed. Also admits YARP source video like input. The module has been tested with handstracking model and ssd_mobilenet_v1_coco. This module also publish detection results in YARP port.
Figure 1. Detection example
tensorflowDetection2D requires a trained model and labels map file. Trained model use .pb extension and labels map file .pbtxt extension. Files shared by Drive.
tensorflowDetection2D use resource finder to locate models and label map files. Set downloaded files in resource finder allowed path.
Allowed paths can be shown:
yarp resource - Create or configure YARP Server.
yarp server
yarp conf pc-ip 10000- Execute programs/tensorflowDetection2D the detector.
cd programs
mkdir build
cd build
cmake ..
make
./tensorflowDetection2D- Connect video source to
tensorflowDetection2D.
yarp connect /videoSource /tensorflowDetection2D/img:iNOTE:
- Video results are published on
/tensorflowDetection2D/img:o - Data results are published on
/tensorflowDetection2D/data:o - Coordinate results are published on
/tensorflowDetection2D/coord:o
tensorflowDetection2D requires:
Tested on: ubuntu 14.04, ubuntu 16.04, ubuntu 18.04 and lubuntu 18.04.
