yolov8 obb_head解析

原创

已于 2024-09-29 20:13:33 修改 · 1.1k 阅读

12 ·

CC 4.0 BY-SA版权

文章标签：

#YOLO #深度学习 #人工智能

于 2024-09-29 13:12:16 首次发布

在学习obb_head之前，请务必理解了hbb_head的原理，然后再来看本文。

class OBB(Detect):
    """YOLOv8 OBB detection head for detection with rotation models."""

    def __init__(self, nc=80, ne=1, ch=()):
        """Initialize OBB with number of classes `nc` and layer channels `ch`."""
        super().__init__(nc, ch)
        self.ne = ne  # number of extra parameters
        self.detect = Detect.forward

        c4 = max(ch[0] // 4, self.ne)
        self.cv4 = nn.ModuleList(nn.Sequential(Conv(x, c4, 3), Conv(c4, c4, 3), nn.Conv2d(c4, self.ne, 1)) for x in ch)

    def forward(self, x):
        """Concatenates and returns predicted bounding boxes and class probabilities."""
        bs = x[0].shape[0]  # batch size
        angle = torch.cat([self.cv4[i](x[i]).view(bs, self.ne, -1) for i in range(self.nl)], 2)  # OBB theta logits
        # NOTE: set `angle` as an attribute so that `decode_bboxes` could use it.
        angle = (angle.sigmoid() - 0.25) * math.pi  # [-pi/4, 3