DTC: A Deformable Transposed Convolution Module for Medical Image Segmentation
This addresses the issue of artifacts and detail loss in medical image segmentation for applications like diagnosis, though it is an incremental improvement over existing deformable convolution methods.
The paper tackled the problem of fixed-position upsampling in medical image segmentation by proposing a Deformable Transposed Convolution (DTC) module, which learns dynamic sampling positions to improve feature reconstruction and detail recovery, as demonstrated on datasets like BTCV15, ISIC18, and BUSI.
In medical image segmentation, particularly in UNet-like architectures, upsampling is primarily used to transform smaller feature maps into larger ones, enabling feature fusion between encoder and decoder features and supporting multi-scale prediction. Conventional upsampling methods, such as transposed convolution and linear interpolation, operate on fixed positions: transposed convolution applies kernel elements to predetermined pixel or voxel locations, while linear interpolation assigns values based on fixed coordinates in the original feature map. These fixed-position approaches may fail to capture structural information beyond predefined sampling positions and can lead to artifacts or loss of detail. Inspired by deformable convolutions, we propose a novel upsampling method, Deformable Transposed Convolution (DTC), which learns dynamic coordinates (i.e., sampling positions) to generate high-resolution feature maps for both 2D and 3D medical image segmentation tasks. Experiments on 3D (e.g., BTCV15) and 2D datasets (e.g., ISIC18, BUSI) demonstrate that DTC can be effectively integrated into existing medical image segmentation models, consistently improving the decoder's feature reconstruction and detail recovery capability.