http://mi.eng.cam.ac.uk/projects/segnet/ A Deep Convolutional Encoder-Decoder Architecture for Robust Semantic Pixel-Wise Labelling Use a random image, upload your own, search for a place, or click on one of the example images in the gallery below. SegNet is trained to lassify each pixel of an urban street image to be one of twelve classes: pedestrian, car, building, tar road, trees etc.
Kai Yan Edit
- http://diydrones.com/profile/yankai implements on Jetson tegra(nvidia) on camera Cctv_cameras#See3CAM(https://www.e-consystems.com/See3CAM-80.asp). See Caffe_berkeley_vision#OpenKAI of vision platform on uav identifying objects on ground(person between cows).