The KITTI Vision Benchmark Suite

The proposed Shared Feature Reuse Segmentation
network has the following novelties: a new yet
effective shared-branch multiple sub-encoders
design, a context mining module and a semantic
aggregating module for better context granularity.
In particular, the shared branches improves the
entire feature hierarchy by sharing the spatial
and context knowledge in both shallow and deep
branches. and the cascading context mining module
filters out the noisy spatial details from the
feature maps and provides a diverse size of
receptive fields for capturing the latent context
between multi-scale geometric shapes in the scene.

SFRSeg has only 1.6 million parameters

@article{SINGHA2023109557,
title = {A real-time semantic segmentation model
using iteratively shared features in multiple sub-
encoders},
journal = {Pattern Recognition},
volume = {140},
pages = {109557},
year = {2023},
issn = {0031-3203},
doi =
{https://doi.org/10.1016/j.patcog.2023.109557},
url =
{https://www.sciencedirect.com/science/article/pii
/S0031320323002571},
author = {Tanmay Singha and Duc-Son Pham and
Aneesh Krishna},
keywords = {Semantic segmentation, Deep
convolution neural networks, Multi-encoder,
Decoder, Feature scaling, Feature aggregation,
Feature reuse, Resource-
}