duoshili2

  视觉目标识别与跟踪是图像分析中的重要研究领域，在现代化社会中具有广泛的发展前景和应用价值。本文深入研究多示例学习方法，将其应用在视觉目标识别与跟踪领域，解决了在目标识别领域中需要大量制作无歧义样本的问题，解决了目标跟踪领域中，易出现跟踪漂移的问题。(Visual object recognition and tracking is an important image analysis research field, has a broad development prospects and value in the modern society. In this paper, in-depth study of multi-instance learning method, its application in the field of visual object recognition and tracking, resolved in the field of object recognition requires a lot of making unambiguous sample problems to solve in the field of target tracking, issue tracking drift prone.)