Application #1: Semantic Segmentation

Env Info:
CentOS/7.2 x86_64 | Server: 2 cores, 8G memory, 5Mbps, No Graphic Card | TensorFlow-2.2.0(stable)
Usage:
Implementation of Deeplabv3+
Image Uploader:

Result:

Application #2: Mask R-CNN

Env Info:
CentOS/7.2 x86_64 | Server: 12 cores, 64G memory, 2Mbps, NVIDIA 1060 6G | TensorFlow-2.2.0(stable)
Usage:
Implementation of MaskRCNN
Image Uploader:

Result:

Application #3: Mask R-CNN and Stereo Vision Based Structural Damage Analysis

Env Info:
CentOS/7.2 x86_64 | Server: 12 cores, 64G memory, 2Mbps, NVIDIA 1060 6G | TensorFlow-2.2.0(stable)
Usage:
Pretrained maskrcnn damage models needed! Training set should include
1.Wall and Ceiling Cracks
2.Uneven Floors
3.Crumbling Concrete
4.Warped Ceilings
This example is using mscoco dataset, stereo images and calibration data is from KITTI dataset.
Since each pair of images calibration data is different in KITTI, only the first validation image is using correct calibration data.
Click to Download Validation dataset
Image Uploader:







Result:

Application #4: Low Network Latency Remote Control

Will be updated in future version.

Application #5: Real Time Online Processing and Analysis From Video Streaming

Will be updated in future version.

Application #6: Real Time Point Cloud Processing from Lidar Streaming Data

Will be updated in future version.

Application #7: Cloud Computing and Stereo Video Streaming Based Autopilot

Will be updated in future version.