Input: pc1: (B,N,C) torch tensor pc2: (B,M,C) torch tensor l1smooth: bool, whether to use l1smooth loss delta: sc

Accumulate one batch of prediction and groundtruth. Args: batch_pred_map_cls: a list of lists [[(pred_cls, pred_box_para

Args: error: Torch tensor (d1,d2,...,dk) Returns: loss: Torch tensor (d1,d2,...,dk) x = error = pred - gt or dist(pred,g

Compute P and R for predicted bounding boxes. Ignores classes! Args: labels: (N x bbox) ground-truth bounding boxes (6 dims) pred

Computes IoU of two axis aligned bboxes. Args: box_a, box_b: 6D of center and lengths Returns: iou

Inverse function to angle2class

Inverse function to size2class

Compute area of two convex hull's intersection area. p1,p2 are a list of (x,y) tuples of hull vertices. return a list of (x,y) for th

Extract scene point clouds and bounding boxes (centroids, box sizes, heading angles, semantic classes). Dumped point clouds and boxes are in

Ref: http://stackoverflow.com/questions/24467972/calculate-area-of-polygon-given-x-y-coordinates

Use accumulated predictions and groundtruths to compute Average Precision.

Flip X-right,Y-forward,Z-up to X-right,Y-down,Z-forward Input and output are both (N,3) array

box_size is array(l,w,h), heading_angle is radius clockwise from pos x axis, center is xyz of box center output (8,3) array for 3D box cornde

Parse groundtruth labels to OBB parameters. Args: end_points: dict {center_label, heading_class_label, heading_residual_

Parse predictions to OBB parameters and suppress overlapping boxes Args: end_points: dict {point_clouds, center, heading

Input: (N,3) Output: (N,2) UV and (N,) depth

Input is NxC, output is num_samplexC

Convert continuous angle to discrete class [optinal] also small regression number from class center angle to current angle.

compute distance between box corners to replace iou Args: crnr1, crnr2: Nx3 points of box corners in camera axis (y points down)

Compute 3D bounding box IoU. Input: corners1: numpy array (8,3), assume up direction is negative Y corners2: numpy array (8,3),

corners: (8,3) no assumption on axis direction

corners: (n,8,3) no assumption on axis direction

Generic functions to compute precision/recall for object detection for a single class. Input: pred: map of {img_id: [(bbo

points are XYZ RGB (RGB in 0-255), semantic label as nyu40 ids, instance label as 1-#instance, box as (cx,cy,cz,dx,dy,dz,semantic_label)

pc: (N,3), box3d: (8,3)

Flip X-right,Y-forward,Z-up to X-right,Y-down,Z-forward Input and output are both (N,3) array

Extract the axis aligned box from a pcl or batch of pcls Args: points: Nx3 points or BxNx3 output is 6 dim: xyz pos of center and

Clip a polygon with another polygon. Ref: https://rosettacode.org/wiki/Sutherland-Hodgman_polygon_clipping#Python Args: subjectPolygon:

Rotation about the z-axis.

Convert 3D box size (l,w,h) to size class and size residual

Numpy function for softmax

Returns a dict with following keys: point_clouds: (N,3+C) center_label: (MAX_NUM_OBJ,3) for GT box center XYZ

Inverse function to angle2class. As ScanNet only has axis-alined boxes so angles are always 0.

Inverse function to size2class

Compute 3D bounding box and semantic classification loss. Args: end_points: dict (read-only) Returns: center_loss h

Compute 3D bounding box and semantic classification loss. Args: end_points: dict (read-only) Returns: center_loss h

Compute objectness loss for the proposals. Args: end_points: dict (read-only) Returns: objectness_loss: scalar Tensor

Compute objectness loss for the proposals. Args: end_points: dict (read-only) Returns: objectness_loss: scalar Tensor

Compute vote loss: Match predicted votes to GT votes. Args: end_points: dict (read-only) Returns: vote_loss: scalar Ten

Compute vote loss: Match predicted votes to GT votes. Args: end_points: dict (read-only) Returns: vote_loss: scalar Tensor

Examine and visualize SUN RGB-D data.

Dump results. Args: end_points: dict {..., pred_mask} pred_mask is a binary mask array of size (batch_size, num_

Generic functions to compute precision/recall for object detection for multiple classes. Input: pred_all: map of {img_id:

Collect 3D bounding box statistics. Used for computing mean box sizes.

Calculate the Intersection over Union (IoU) of two 2D bounding boxes. Parameters ---------- bb1 : dict Keys: {'x1', 'x2', 'y

Loss functions Args: end_points: dict { seed_xyz, seed_inds, vote_xyz, center,

Convert predictions to OBB parameters (eight corner points) Args: end_points: dict {center_label, heading_class_label, headin

Log a list of images.