Add image prediction function to TRec models.

fit/modules/TRecTransformerModule.py

@@ -12,7 +12,7 @@
 from torch.nn import functional as F
 import torch.fft
 
-from fit.utils.utils import denormalize, denormalize_amp, denormalize_phi
+from fit.utils.utils import denormalize, denormalize_amp, denormalize_phi, denormalize_FC
 
 
 class TRecTransformerModule(LightningModule):
@@ -27,7 +27,6 @@ def __init__(self, d_model, sinogram_coords, target_coords, src_flatten_coords,
                  attention_type="linear", n_layers=4, n_heads=4, d_query=4, dropout=0.1, attention_dropout=0.1,
                  only_FBP=False,
                  only_convblock=False,
-                 no_convblock=False,
                  d_conv=8):
         super().__init__()
 
@@ -46,7 +45,8 @@ def __init__(self, d_model, sinogram_coords, target_coords, src_flatten_coords,
                                   "d_query",
                                   "dropout",
                                   "attention_dropout",
-                                  "convblock_only",
+                                  "only_FBP",
+                                  "only_convblock",
                                   "d_conv")
         self.sinogram_coords = sinogram_coords
         self.target_coords = target_coords
@@ -145,7 +145,7 @@ def _fc_loss(self, pred_fc, target_fc, amp_min, amp_max):
         return torch.mean(amp_loss + phi_loss), torch.mean(amp_loss), torch.mean(phi_loss)
 
     def criterion(self, pred_fc, pred_img, target_fc, amp_min, amp_max):
-        if self.hparams.convblock_only:
+        if self.hparams.only_convblock:
             return self._real_loss(pred_img=pred_img, target_fc=target_fc, amp_min=amp_min,
                                    amp_max=amp_max), torch.tensor(0.0), torch.tensor(0.0)
         else:
@@ -337,3 +337,28 @@ def test_epoch_end(self, outputs):
         self.log('Mean PSNR', torch.mean(outputs).detach().cpu().numpy(), logger=True)
         self.log('SEM PSNR', torch.std(outputs / np.sqrt(len(outputs))).detach().cpu().numpy(),
                  logger=True)
+
+    def get_imgs(self, x, fbp, y, amp_min, amp_max):
+        self.eval()
+
+        self.bin_factor = 1
+        self.register_buffer('mask', psf_rfft(self.bin_factor, pixel_res=self.hparams.img_shape).to(self.device))
+
+        x_fc_, fbp_fc_, y_fc_ = self._bin_data(x, fbp, y)
+
+        pred_fc, pred_img = self.trec.forward(x_fc_, fbp_fc_, amp_min=amp_min, amp_max=amp_max,
+                                              dst_flatten_coords=self.dst_flatten_order,
+                                              img_shape=self.hparams.img_shape,
+                                              attenuation=self.mask)
+
+        tmp = denormalize_FC(pred_fc, amp_min=amp_min, amp_max=amp_max)
+        pred_fc_ = torch.ones(x.shape[0], self.hparams.img_shape * (self.hparams.img_shape // 2 + 1), dtype=x.dtype,
+                              device=x.device)
+        pred_fc_[:, :tmp.shape[1]] = tmp
+
+        dft_pred_fc = convert2DFT(x=pred_fc, amp_min=amp_min, amp_max=amp_max,
+                                  dst_flatten_order=self.dst_flatten_order, img_shape=self.hparams.img_shape)
+        img_pred_before_conv = torch.roll(torch.fft.irfftn(dft_pred_fc, dim=[1, 2], s=2 * (self.hparams.img_shape,)),
+                                          2 * (self.hparams.img_shape // 2,), (1, 2))
+
+        return pred_img, img_pred_before_conv