Machine learning from previous examples or knowledge is a key element in many image processing and pattern recognition tasks, e.g. clustering, segmentation, stereo matching, optical flow, tracking and object recognition. Acquiring that knowledge frequently requires human labeling of large data sets, which can be difficult and time-consuming to obtain. One way to ameliorate this task is to use Semi-supervised Learning (SSL), which combines both labeled and raw data and incorporates both global consistency (points in the same cluster are likely to have the same label) and local smoothness (nearby points are likely to have the same label). There are a number of vision tasks that can be solved efficiently and accurately using SSL. SSL has been applied extensively in clustering and image segmentation. In this dissertation, we will show that it is also suitable for stereo matching, optical flow and tracking problems. Our novel algorithm has converted the stereo matching problem into a multi-label semi-supervised learning one. It is similar to a diffusion process, and we will show our approach has a closed-form solution for the multi-label problem. It sparks a new direction from the traditional energy minimization approach, such as Graph Cut or Belief Propagation. The occlusion area is detected using the matching confidence level, and solved with local fitting. Our results have been applied in the Middlebury Stereo database, and are within the top 20 best results in terms of accuracy and is considerably faster than the competing approaches. We have also adapted our algorithm, and demonstrated its performance on optical flow problems. Again, our results are compared with the ground truth and state of the art on the Middlebury Flow database, and its advantages in accuracy as well as speed are demonstrated. The above algorithm is also being used in our current NSF sponsored project, an Automated, Real-Time Identification and Monitoring Instrument for Reef Fish Communities, whose goal is to track and recognize tropical fish, initially in an aquarium and ultimately on a coral reef. Our approach, which combines background subtraction and optical flow, automatically finds the correct outline of multiple fish species in the field of view, and tracks the contour reliably over consecutive frames. Currently, near real-time results are being achieved, with a processing frame rate of 3-5 fps. The recent progress in semi-supervised learning applied to image segmentation is also briefly reviewed.