quad-iou 0.0.0

A CUDA extension for calculating IoU(Intersection over Union) for quadrilaterals(MxN) bound to PyTorch(usable with torch tensors).

IoU(Intersection over Union) Calculation for Quadrilaterals(CUDA Kernel)

Cuda kernel for calculating IoU(intersection over union) for quadrilaterals. It can calculate IoU either for 1->1 match or N->M match, returning an iou matrix with N rows and M columns. Torch CUDA extensions are used for running the compiled kernels.

Example usage:

import torch
import quad_iou

# NxM quadrilaterals
a = torch.rand((200, 4, 2)).cuda()
b = torch.rand((300, 4, 2)).cuda()

# sort_input_quads indicate whether kernel should sort the quadrilateral corners
# clockwise before calculating iou
iou_matrix = quad_iou.calculate_iou(a, b, sort_input_quads=True) # returns tensor of shape [200, 300]

# 1x1 case
a = torch.tensor([0.0, 0, 300, 0, 300, 300, 0, 300]).cuda()
b = torch.tensor([0.0, 0, 150, 0, 150, 150, 0, 150]).cuda()
# Module expects tensor of shape [N, 4, 2], so we reshape the tensors
a = a.reshape(-1, 4, 2)
b = b.reshape(-1, 4, 2)
iou = quad_iou.calculate_iou(a, b, sort_input_quads=True)

Comparison with Shapely library

While CPUs and GPUs are not to be compared for speed, we provide a script to demonstrate the potential speedup when using a GPU. To compare the execution time of calculating the IoU for quadrilaterals using the Shapely library versus our GPU-accelerated implementation, run python tryout_scripts/comparison.py