简介:YOLOv11改进策略【损失函数篇】| 引入Soft-NMS,提升密集遮挡场景检测精度,包括GIoU-NMS、DIoU-NMS、CIoU-NMS、SIoU-NMS、 EIou-NMS 一、背景: 传统的非极大值抑制(NMS)算法在目标检测中存在一个问题,即当一个物体的检测框与具有最高得分的检测框M有重叠(在预定义的重叠阈值内)时,会将该检测
因此,本研究旨在改进Soft-nms算法,以提高手势识别和游戏AI对战系统的准确性和响应速度。具体而言,我们将引入一系列新的IoU(Intersection over Union)计算方法,包括GIoU(Generalized IoU)、DIoU(Distance IoU)、CIoU(Complete IoU)、EIoU(Enhanced IoU)、SIoU(Soft IoU)等。这些新的IoU计算方法可以更准确地衡量目标之间...
Soft NMS defpy_cpu_softnms(dets,sc,Nt=0.3,sigma=0.5,thresh=0.001,method=2):"""py_cpu_softnms:param dets: boexs坐标矩阵format [y1, x1, y2, x2]:param sc: 每个 boxes 对应的分数:param Nt: iou 交叠门限:param sigma: 使用 gaussian 函数的方差:param thresh: 最后的分数门限:param method...
Cai, G., Ciou, J.-H., Liu, Y., Jiang, Y. & Lee, P. S. Leaf-inspired multiresponsive MXene-based actuator for programmable smart devices. Sci. Adv. 5, eaaw7956 (2019). Article CAS Google Scholar Li, J. et al. Photothermal bimorph actuators with in-built cooler for light mill...
Soft robotic technologies hold great promises for biomedical applications including wearable, prosthetic robots, and miniaturized surgical devices.
doi:10.1038/s41467-023-37535-4Dace GaoGurunathan ThangavelJunwoo LeeJian LvYi LiJinghao CiouJiaqing XiongT. ParkPooi See LeeNature Publishing Group UKNature Communications
主要不同就是将iou换成了diou ,使用的还是同一个函数,只是改变了一下参数,所以其实还可以使用giou_nms与ciou_nms,本质上没有变化。 4. NMS变体代码完整展示 需要注意,以下代码和yolov3spp代码是不一样的,不过可以直接替换使用。yolov3spp中使用的方法只是hard_nms处理,并且设置了一个可控参数选择是否使用merge_...
My implementation of label-smooth, amsoftmax, partial-fc, focal-loss, dual-focal-loss, triplet-loss, giou/diou/ciou-loss/func, affinity-loss, pc_softmax_cross_entropy, ohem-loss(softmax based on line hard mining loss), large-margin-softmax(bmvc2019), lovasz-softmax-loss, and dice-loss...
Cai, G., Ciou, J.-H., Liu, Y., Jiang, Y. & Lee, P. S. Leaf-inspired multiresponsive MXene-based actuator for programmable smart devices.Sci. Adv.5, eaaw7956 (2019). ArticleCASGoogle Scholar Li, J. et al. Photothermal bimorph actuators with in-built cooler for light mills, freq...
def nms_iou(box1, box2, GIoU=False, DIoU=False, CIoU=False, SIoU=False, EIou=False, eps=1e-7): # Returns Intersection over Union (IoU) of box1(1,4) to box2(n,4) b1_x1, b1_y1, b1_x2, b1_y2 = box1.chunk(4, -1) b2_x1, b2_y1, b2_x2, b2_y2 = box2.chunk(...