YOLO9000: Better, Faster, Stronger

PaperReading 2017.12 ObjectDetection

• Motivation

  • General purpose object detection should be fast, accurate, and able to recognize a wide variety of objects.

  • We also propose a joint training algorithm that allows us to train object detectors on both detection and classification data.

• Better

  • batch normalization

  • high resolution

    增大了输入图片的大小

  • convolutional with anchor boxes

  • dimension clusters

  • direct location prediction

  • fine-grained features

  • multi-scale training

    多尺度的输入

  • further experiment

• Faster

  • Darknet-19

  • training for

• Stronger

  • hierarchical classification

    利用了wordTree的启发信息

  • Dataset combination with WordTree

  • Joint classification and detection

• Conclusion