From a15d2d27e9a21cd5f9885a6c509099fdab33e559 Mon Sep 17 00:00:00 2001
From: Andre Anjos <andre.anjos@idiap.ch>
Date: Mon, 29 Jun 2020 14:53:44 +0200
Subject: [PATCH] [engine.evaluator] Fix function to calculate patch
performance
---
bob/ip/binseg/engine/evaluator.py | 35 ++++++++++++++++++++++++-------
1 file changed, 28 insertions(+), 7 deletions(-)
diff --git a/bob/ip/binseg/engine/evaluator.py b/bob/ip/binseg/engine/evaluator.py
index 1b5be9f7..37fb4e2b 100644
--- a/bob/ip/binseg/engine/evaluator.py
+++ b/bob/ip/binseg/engine/evaluator.py
@@ -4,6 +4,7 @@
"""Defines functionality for the evaluation of predictions"""
import os
+import itertools
import PIL
import numpy
@@ -11,6 +12,7 @@ import pandas
from tqdm import tqdm
import torch
+import torch.nn.functional
import torchvision.transforms.functional as VF
import h5py
@@ -184,13 +186,32 @@ def _patch_measures(pred, gt, steps, size):
"""
- height, width, stride = window_size
- pred_patches = pred.unfold(0, height, stride).unfold(1, width, stride)
- gt_patches = unfold(0, height, stride).unfold(1, width, stride)
-
- # add patch number for each set of measures
- dfs = [_sample_measures(p, g, step) for p,g in zip(pred_patches, gt_patches)]
- for i, k in enumerate(dfs): k['patch'] = i
+ height, width, stride = size
+
+ # we calculate the required padding so that the last windows on the left
+ # and bottom size of predictions/ground-truth data are zero padded, and
+ # torch unfolding works exactly.
+ padding = (0, 0)
+ rem = (pred.shape[1] - width) % stride
+ if rem != 0:
+ padding = (0, (stride-rem))
+ rem = (pred.shape[0] - height) % stride
+ if rem != 0:
+ padding += (0, (stride-rem))
+
+ pred_padded = torch.nn.functional.pad(pred, padding)
+ gt_padded = torch.nn.functional.pad(gt.squeeze(0), padding)
+
+ # this will create as many views as required
+ pred_patches = pred_padded.unfold(0, height, stride).unfold(1, width, stride)
+ gt_patches = gt_padded.unfold(0, height, stride).unfold(1, width, stride)
+ assert pred_patches.shape == gt_patches.shape
+ ylen, xlen, _, _ = pred_patches.shape
+
+ dfs = []
+ for j, i in itertools.product(range(ylen), range(xlen)):
+ dfs.append(_sample_measures(pred_patches[j,i,:,:], gt_patches[j,i,:,:], steps))
+ dfs[-1]['patch'] = i+(j*xlen)
return pandas.concat(dfs, ignore_index=True)
--
GitLab