News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
Target classification in AI computer vision
Target classification in AI computer vision
One, the classic task
There are three classical tasks related to targets in the field of computer vision: classification, detection and segmentation. The classification is to tell you what is "what". The goal of the next two tasks is to tell you where "where", and the division task will answer this question at the pixel level.
It
Two, the meaning of several professional nouns
Target detection, searching for the target area of the system.
Target tracking, capturing the trajectory of the region of interest.
The target classification is divided into human, vehicle or other moving objects.
Target behavior recognition, behavior recognition of tracking targets.
Stereo vision matching is a technology for recovering depth information from planar images.
Optical flow refers to the surface movement of grayscale pattern and the projection of 3D motion field on two-dimensional image plane.
Scene flow is the three-dimensional motion field formed by scene movement in space. They are widely used in video tracking and monitoring, autonomous robot navigation, 3D video compression and display.
Three, KITTI data set (very important)
For the KITTI data set, the data set is used to evaluate the performance of computer vision technology such as stereo, optical flow, visual odometry, 3D object detection (object detection) and 3D tracking (tracking) in the vehicle environment.
Four, several kinds of deep learning algorithms
FCNN: full convolution network: parallel, iteration
Sharing space structure parameters on the basis of CNN:FCNN
Sharing time structure parameters on the basis of RNN:FCNN
There are three classical tasks related to targets in the field of computer vision: classification, detection and segmentation. The classification is to tell you what is "what". The goal of the next two tasks is to tell you where "where", and the division task will answer this question at the pixel level.
It
Two, the meaning of several professional nouns
Target detection, searching for the target area of the system.
Target tracking, capturing the trajectory of the region of interest.
The target classification is divided into human, vehicle or other moving objects.
Target behavior recognition, behavior recognition of tracking targets.
Stereo vision matching is a technology for recovering depth information from planar images.
Optical flow refers to the surface movement of grayscale pattern and the projection of 3D motion field on two-dimensional image plane.
Scene flow is the three-dimensional motion field formed by scene movement in space. They are widely used in video tracking and monitoring, autonomous robot navigation, 3D video compression and display.
Three, KITTI data set (very important)
For the KITTI data set, the data set is used to evaluate the performance of computer vision technology such as stereo, optical flow, visual odometry, 3D object detection (object detection) and 3D tracking (tracking) in the vehicle environment.
Four, several kinds of deep learning algorithms
FCNN: full convolution network: parallel, iteration
Sharing space structure parameters on the basis of CNN:FCNN
Sharing time structure parameters on the basis of RNN:FCNN