4.2. Extended-Efficient Layer Aggregation Networks Module

This model introduces the Extended-efficient layer aggregation networks (E-ELAN) module, elevating network learning potential through the Expand, Shuffle, Merge Cardinality Network (EALN) approach. The E-ELAN module modifies both the backbone network and the head network’s structure [38]. Group convolution is employed to expand the feature base count, and features from different groups are fused using shuffling and merging cardinality operations. This strategy improves parameter utilization, computational efficiency, and features learned from various feature maps. Figure 3 illustrates the architecture of the E-ELAN module.

E-ELAN module’s architectural depiction.