Extract filtered in-range velodyne coordinates based on azimuth & elevation angle limit Args: `x`:velodyne points x array `y`:velodyne points y

Computes pairwise intersection-over-union between box collections. Args: boxes1: a numpy array with shape [N, 4] holding N boxes.

calculate box iou. note that jit version runs 2x faster than cython in my machine! Parameters ---------- boxes: (N, 4) ndarray of

nms in gpu. Args: dets ([type]): [description] nms_overlap_thresh ([type]): [description] device_id ([type], op

convert kitti points(N, 4) to a bev map. return [C, H, W] map. this function based on algorithm in points_to_voxel. takes 5ms in a reduced poi

get cls_weights and reg_weights from labels.

Return quaternion from Euler angles and axis sequence. ai, aj, ak : Euler's roll, pitch and yaw angles axes : One of 24 axis sequences as str

Return quaternion from Euler angles and axis sequence. ai, aj, ak : Euler's roll, pitch and yaw angles axes : One of 24 axis sequences as str

Return quaternion from Euler angles and axis sequence. ai, aj, ak : Euler's roll, pitch and yaw angles axes : One of 24 axis sequences as str

Parse lines in 2D detection output files

Parse lines in 2D detection output files

Converts an RGB list to the packed float format used by PCL From the PCL docs: "Due to historical reasons (PCL was first develop

rotation 2d points based on origin point clockwise when angle positive. Args: points (float array, shape=[N, point_size, 2]): poin

rotation 2d points based on origin point clockwise when angle positive. Args: points (float array, shape=[N, point_size, 2]): points

save a model into model_dir. Args: model_dir: string, indicate your model dir(save ckpts, summarys, logs, etc). model:

pytorch edition of tensorflow scatter_nd. this function don't contain except handle code. so use this carefully when indice repeats, don't sup