# Explicit Over Implicit: Enhancing CNNs Via Complex Structure Tensor Representations for Periocular Recognition

arXiv: https://arxiv.org/abs/2607.15410 · HTML: https://mlsanity.com/paper/2607.15410/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 46 references that are in the map.
- First posted on arXiv 16 Jul 2026.

## Builds on (most cited first, 8 of 8)

- [ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness](https://mlsanity.com/paper/1811.12231/) (2018)
- [Local Binary Convolutional Neural Networks](https://mlsanity.com/paper/1608.06049/) (2016)
- [A Survey on Periocular Biometrics Research](https://mlsanity.com/paper/1810.03360/) (2018)
- [GaborNet: Gabor filters with learnable parameters in deep convolutional neural networks](https://mlsanity.com/paper/1904.13204/) (2019)
- [Near-infrared and visible-light periocular recognition with Gabor features using frequency-adaptive automatic eye detection](https://mlsanity.com/paper/2211.05544/) (2022)
- [Deep Representations for Cross-spectral Ocular Biometrics](https://mlsanity.com/paper/1911.09509/) (2019)
- [One-Shot Learning for Periocular Recognition: Exploring the Effect of Domain Adaptation and Data Bias on Deep Representations](https://mlsanity.com/paper/2307.05128/) (2023)
- [Experimental analysis regarding the influence of iris segmentation on the recognition rate](https://mlsanity.com/paper/2211.05507/) (2022)

## Abstract (arXiv)

Our study provides evidence that CNNs struggle to extract orientation features effectively. We show that using the Complex Structure Tensor, which contains compact orientation features with certainties, as input to CNNs consistently improves identification accuracy compared to grayscale inputs alone. Experiments also demonstrated that our inputs, provided by mini-complex convnets, combined with reduced CNN sizes, outperformed full-fledged, prevailing CNN architectures. This suggests that the upfront use of orientation features in CNNs, a strategy seen in mammalian vision, not only mitigates their limitations but also enhances their explainability and relevance to thin-clients. Experiments were conducted on publicly available datasets comprising periocular images (Cross-Eyed and PolyU) for biometric identification and verification in both Close-World and Open-World Scenarios using six CNN architectures. Our experiments on the Cross-Eyed and PolyU datasets yield a 5-26% reduction in EER, providing strong empirical evidence that explicit orientation priors mitigate CNN representational limits in Open-World and Close-World scenarios.