mihaidusmanu

# python venv
VENV_PATH=/home/dusmanum/.venvs
function create_venv () {
  if [[ $# -ne 1 ]]; then
    echo "Error: Expected env name as parameter"
    return
  fi
  echo "Creating env $1 in $VENV_PATH/$1"
  python -m venv $VENV_PATH/$1
}

mihaidusmanu

from pathlib import Path
from scantools.capture import Capture

path = Path('/media/SSD2/ms_eth_dataset/datasets/release_public/v2.0')

for location in ['LIN', 'CAB', 'HGE']:
    print(location)
    capture = Capture.load(path / location)
 
    for session_id in ['map', 'query_hololens', 'query_phone', 'query_val_hololens', 'query_val_phone']:

mihaidusmanu

\RequirePackage[%
  style         = alphabetic,
  autocite      = inline,
  backend       = biber,
  doi           = false,
  url           = true,
  giveninits    = true,
  hyperref      = true,
  sortcites     = true,
  minalphanames = 1,

mihaidusmanu

import torch
import torch.nn as nn
import torch.nn.functional as F


class GeM(nn.Module):
    def __init__(self, p=3, eps=1e-6):
        # Based on https://arxiv.org/abs/1711.02512.
        super().__init__()
        self.p = torch.nn.parameter.Parameter(torch.ones(1) * p)

mihaidusmanu

def compute_reconstruction_statistics(reference_model_path):
    # Images w. intrinsics and extrinsics.
    with open(os.path.join(reference_model_path, 'cameras.txt'), 'r') as f:
        raw_cameras = f.readlines()[3 :]

    cameras = {}
    for raw_line in raw_cameras:
        split_line = raw_line.strip('\n').split(' ')
        cameras[int(split_line[0])] = split_line[1 :]

mihaidusmanu

Compiling COLMAP on Leonhard

This is a short tutorial for compiling COLMAP on Leonhard with GCC 6.3.0. Qt installation is quite large so COLMAP cannot be compiled in the home directories.

Modules and Paths

Create a file called colmap_startup.sh in your home directory containing the following:

export DATA=/cluster/project/infk/cvg/$USER

mihaidusmanu

import numpy as np

def mutual_nn_matcher(descriptors1, descriptors2):
    descriptors1 = descriptors1 / np.linalg.norm(descriptors1, axis=1)[:, np.newaxis]
    descriptors2 = descriptors2 / np.linalg.norm(descriptors2, axis=1)[:, np.newaxis]
    similarity = descriptors1 @ descriptors2.transpose()
    nn12 = np.argmax(similarity, axis=1)
    nn21 = np.argmax(similarity, axis=0)
    ids1 = np.arange(similarity.shape[0])
    valid_matches = (ids1 == nn21[nn12])

mihaidusmanu

mihaidusmanu

mihaidusmanu

from keras.engine.topology import Layer

import numpy as np

import tensorflow as tf

class MaxPooling2D(Layer):
  def __init__(self, pool_size = 2, stride = None, padding = 'VALID', **kwargs):
    self.pool_size = pool_size
    assert(isinstance(self.pool_size, int))