Pawel Piwowarski
init commit
0a82b18
"""
"XFeat: Accelerated Features for Lightweight Image Matching, CVPR 2024."
https://www.verlab.dcc.ufmg.br/descriptors/xfeat_cvpr24/
Camera pose metrics adapted from LoFTR https://github.com/zju3dv/LoFTR/blob/master/src/utils/metrics.py
The main difference is the use of poselib instead of OpenCV's vanilla RANSAC for E_mat, which is more stable and MUCH and faster.
"""
import argparse, glob, sys, os, time
import torch
from torch.utils.data import Dataset, DataLoader
import cv2
import numpy as np
import poselib
import json
import copy
import tqdm
# Disable scientific notation
np.set_printoptions(suppress=True)
class MegaDepth1500(Dataset):
"""
Streamlined MegaDepth-1500 dataloader. The camera poses & metadata are stored in a formatted json for facilitating
the download of the dataset and to keep the setup as simple as possible.
"""
def __init__(self, json_file, root_dir):
# Load the info & calibration from the JSON
with open(json_file, 'r') as f:
self.data = json.load(f)
self.root_dir = root_dir
if not os.path.exists(self.root_dir):
raise RuntimeError(
f"Dataset {self.root_dir} does not exist! \n \
> If you didn't download the dataset, use the downloader tool: python3 -m modules.dataset.download -h")
def __len__(self):
return len(self.data)
def __getitem__(self, idx):
data = copy.deepcopy(self.data[idx])
h1, w1 = data['size0_hw']
h2, w2 = data['size1_hw']
# Here we resize the images to max_dim = 1200, as described in the paper, and adjust the image such that it is divisible by 32
# following the protocol of the LoFTR's Dataloader (intrinsics are corrected accordingly).
# For adapting this with different resolution, you would need to re-scale intrinsics below.
image0 = cv2.resize( cv2.imread(f"{self.root_dir}/{data['pair_names'][0]}"),
(w1, h1))
image1 = cv2.resize( cv2.imread(f"{self.root_dir}/{data['pair_names'][1]}"),
(w2, h2))
data['image0'] = torch.tensor(image0.astype(np.float32)/255).permute(2,0,1)
data['image1'] = torch.tensor(image1.astype(np.float32)/255).permute(2,0,1)
for k,v in data.items():
if k not in ('dataset_name', 'scene_id', 'pair_id', 'pair_names', 'size0_hw', 'size1_hw', 'image0', 'image1'):
data[k] = torch.tensor(np.array(v, dtype=np.float32))
return data
################################# Metrics #####################################
def relative_pose_error(T_0to1, R, t, ignore_gt_t_thr=0.0):
# angle error between 2 vectors
t_gt = T_0to1[:3, 3]
n = np.linalg.norm(t) * np.linalg.norm(t_gt)
t_err = np.rad2deg(np.arccos(np.clip(np.dot(t, t_gt) / n, -1.0, 1.0)))
t_err = np.minimum(t_err, 180 - t_err) # handle E ambiguity
if np.linalg.norm(t_gt) < ignore_gt_t_thr: # pure rotation is challenging
t_err = 0
# angle error between 2 rotation matrices
R_gt = T_0to1[:3, :3]
cos = (np.trace(np.dot(R.T, R_gt)) - 1) / 2
cos = np.clip(cos, -1., 1.) # handle numercial errors
R_err = np.rad2deg(np.abs(np.arccos(cos)))
return t_err, R_err
def intrinsics_to_camera(K):
px, py = K[0, 2], K[1, 2]
fx, fy = K[0, 0], K[1, 1]
return {
"model": "PINHOLE",
"width": int(2 * px),
"height": int(2 * py),
"params": [fx, fy, px, py],
}
def estimate_pose_poselib(kpts0, kpts1, K0, K1, thresh, conf=0.99999):
M, info = poselib.estimate_relative_pose(
kpts0, kpts1,
intrinsics_to_camera(K0),
intrinsics_to_camera(K1),
{"max_epipolar_error": thresh,
"success_prob": conf,
"min_iterations": 20,
"max_iterations": 1_000},
)
R, t, inl = M.R, M.t, info["inliers"]
inl = np.array(inl)
ret = (R, t, inl)
return ret, (kpts0, kpts1)
def tensor2bgr(t):
return (t.cpu()[0].permute(1,2,0).numpy()*255).astype(np.uint8)
def compute_pose_error(pair):
"""
Input:
pair (dict):{
"pts0": ndrray(N,2)
"pts1": ndrray(N,2)
"K0": ndrray(3,3)
"K1": ndrray(3,3)
"T_0to1": ndrray(4,4)
}
Update:
pair (dict):{
"R_err" List[float]: [N]
"t_err" List[float]: [N]
"inliers" List[np.ndarray]: [N]
}
"""
pixel_thr = 1.0 if 'ransac_thr' not in pair else pair['ransac_thr']
conf = 0.99999
pair.update({'R_err': np.inf, 't_err': np.inf, 'inliers': []})
pts0 = pair['pts0']
pts1 = pair['pts1']
K0 = pair['K0'].cpu().numpy()[0]
K1 = pair['K1'].cpu().numpy()[0]
T_0to1 = pair['T_0to1'].cpu().numpy()[0]
ret, corrs = estimate_pose_poselib(pts0, pts1, K0, K1, pixel_thr, conf=conf)
if ret is not None:
R, t, inliers = ret
t_err, R_err = relative_pose_error(T_0to1, R, t, ignore_gt_t_thr=0.0)
pair['R_err'] = R_err
pair['t_err'] = t_err
def error_auc(errors, thresholds=[5, 10, 20]):
"""
Args:
errors (list): [N,]
thresholds (list)
"""
errors = [0] + sorted(list(errors))
recall = list(np.linspace(0, 1, len(errors)))
aucs = []
for thr in thresholds:
last_index = np.searchsorted(errors, thr)
y = recall[:last_index] + [recall[last_index-1]]
x = errors[:last_index] + [thr]
aucs.append(np.trapz(y, x) / thr)
return {f'auc@{t}': auc for t, auc in zip(thresholds, aucs)}
def compute_maa(pairs, thresholds=[5, 10, 20]):
print("auc / mAcc on %d pairs" % (len(pairs)))
errors = []
for p in pairs:
et = p['t_err']
er = p['R_err']
errors.append(max(et, er))
d_err_auc = error_auc(errors)
for k,v in d_err_auc.items():
print(k, ': ', '%.1f'%(v*100))
errors = np.array(errors)
for t in thresholds:
acc = (errors <= t).sum() / len(errors)
print("mAcc@%d: %.1f "%(t, acc*100))
@torch.inference_mode()
def run_pose_benchmark(matcher_fn, loader, ransac_thr=2.5):
"""
Run relative pose estimation benchmark using a specified matcher function and data loader.
Parameters
----------
matcher_fn : callable
The matching function to be evaluated for pose estimation. It should accept two np.array RGB images (H,W,3)
and return mkpts_0, mkpts_1 which are np.array(N,2) matching coordinates.
loader : iterable
Data loader that provides batches of data. Each batch should contain two images, along
with their groundtruth camera poses.
ransac_thr : float, optional, default=2.5
The RANSAC threshold for considering a point as an inlier in pixels.
"""
pairs = []
cnt = 0
for d in tqdm.tqdm(loader):
d_error = {}
src_pts, dst_pts = matcher_fn(tensor2bgr(d['image0']), tensor2bgr(d['image1']))
#delete images to avoid OOM, happens in low mem machines
del d['image0']
del d['image1']
#rescale kpts
src_pts = src_pts * d['scale0'].numpy()
dst_pts = dst_pts * d['scale1'].numpy()
d.update({"pts0":src_pts, "pts1": dst_pts,'ransac_thr': ransac_thr})
compute_pose_error(d)
pairs.append(d)
cnt+=1
compute_maa(pairs)
def parse_args():
parser = argparse.ArgumentParser(description="Run pose benchmark with matcher")
parser.add_argument('--dataset-dir', type=str, required=True,
help="Path to MegaDepth dataset root")
parser.add_argument('--matcher', type=str, choices=['xfeat', 'xfeat-star', 'alike'], default='xfeat',
help="Matcher to use (xfeat or alike)")
parser.add_argument('--ransac-thr', type=float, default=2.5,
help="RANSAC threshold value in pixels (default: 2.5)")
return parser.parse_args()
if __name__ == '__main__':
args = parse_args()
dataset = MegaDepth1500( json_file = './assets/megadepth_1500.json',
root_dir = args.dataset_dir + "/megadepth_test_1500")
loader = DataLoader(dataset, batch_size=1, shuffle=False)
if args.matcher == 'xfeat':
print("Running benchmark for XFeat..")
from modules.xfeat import XFeat
xfeat = XFeat()
run_pose_benchmark(matcher_fn = xfeat.match_xfeat, loader = loader, ransac_thr = args.ransac_thr)
elif args.matcher == 'xfeat-star':
from modules.xfeat import XFeat
print("Running benchmark for XFeat*..")
xfeat = XFeat(top_k = 10_000)
run_pose_benchmark(matcher_fn = xfeat.match_xfeat_star, loader = loader, ransac_thr = args.ransac_thr)
elif args.matcher == 'alike':
from third_party import alike_wrapper as alike
print("Running benchmark for ALIKE..")
run_pose_benchmark(matcher_fn = alike.match_alike, loader = loader, ransac_thr = args.ransac_thr)