SIPFormer: Segmentation of Multiocular Biometric Traits With Transformers

Hassan, Bilal; Hassan, Taimur; Ahmed, Ramsha; Werghi, Naoufel; Dias, Jorge

doi:10.1109/tim.2022.3232162

Cited by 5 publications

(3 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Incorporating SA into the network can help it focus more on the defective feature data without adding too much computation. It can also help obtain better image details and improve the output quality [27].…”

Section: Image Enhancement Methods For Flexible Packaging Shape Defectsmentioning

confidence: 99%

An online color and shape integrated detection method for flexible packaging surface defects

Sun,

Wei,

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

It is difficult for the spectrophotometer to meet the requirement of real-time color defect detection for flexible packaging prints. The false of shape defect detection is caused by artifact interference and insufficient classification accuracy of defect classification network. A color defect detection method for flexible packaging is proposed, which realizes the adaptive adjustment of the correction parameters of the Commission Internationale de l´Eclairage Delta E 2000 (CIEDE2000) equations with the detection object. It improves the speed and accuracy of the color defect detection for flexible packaging. An quadratic difference strategy is designed for template matching subtraction method to remove artifact interference. A method for enhancing shape defect data set of flexible packaging is proposed. Using discrete images of defects as network input, self-attention (SA) mechanism and spectral normalization (SN) methods are added to the deep convolutional generative adversarial networks (DCGAN) to enhance the effective dataset for the training of defect classification network. The accuracy of color detection for flexible packaging prints is improved by 38.7% based on optimized CIEDE2000 The average structure similarity index measure (SSIM) value of the improved DCGAN for defect detection is 0.845, and the Fréchet inception distance (FID) is 121.463. It takes 83.63 ms for the color and shape integrated detection method to detect shape defects on flexible packaging surfaces with an accuracy of 98.3%. The online color and shape integrated detection method can be applied to automated flexible packaging workshops to achieve real-time defect detection.

show abstract

Section: Image Enhancement Methods For Flexible Packaging Shape Defectsmentioning

confidence: 99%

An online color and shape integrated detection method for flexible packaging surface defects

Sun,

Wei,

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Hassan et al [19] introduced SIPFormer, a novel framework comprising encoder, decoder, and transformer modules designed for joint segmentation. Their approach includes a pre-processing stage to enhance eye features while suppressing information from the periocular regions.…”

Section: Related Workmentioning

confidence: 99%

“…In recent years, to enhance segmentation performance with limited annotated datasets, methods based on semi-supervised learning have been proposed [18]. Hassan et al introduced a new framework named SIPFormer [19], which integrates transformer architecture for this multi-modal segmentation task. At the same time, it improves segmentation accuracy and introduces many parameters, reducing model efficiency.…”

Section: Introductionmentioning

confidence: 99%

OcularSeg: Accurate and Efficient Multi-Modal Ocular Segmentation in Non-Constrained Scenarios

Zhang,

Wang,

et al. 2024

Electronics

View full text Add to dashboard Cite

Multi-modal ocular biometrics has recently garnered significant attention due to its potential in enhancing the security and reliability of biometric identification systems in non-constrained scenarios. However, accurately and efficiently segmenting multi-modal ocular traits (periocular, sclera, iris, and pupil) remains challenging due to noise interference or environmental changes, such as specular reflection, gaze deviation, blur, occlusions from eyelid/eyelash/glasses, and illumination/spectrum/sensor variations. To address these challenges, we propose OcularSeg, a densely connected encoder–decoder model incorporating eye shape prior. The model utilizes Efficientnetv2 as a lightweight backbone in the encoder for extracting multi-level visual features while minimizing network parameters. Moreover, we introduce the Expectation–Maximization attention (EMA) unit to progressively refine the model’s attention and roughly aggregate features from each ocular modality. In the decoder, we design a bottom-up dense subtraction module (DSM) to amplify information disparity between encoder layers, facilitating the acquisition of high-level semantic detailed features at varying scales, thereby enhancing the precision of detailed ocular region prediction. Additionally, boundary- and semantic-guided eye shape priors are integrated as auxiliary supervision during training to optimize the position, shape, and internal topological structure of segmentation results. Due to the scarcity of datasets with multi-modal ocular segmentation annotations, we manually annotated three challenging eye datasets captured in near-infrared and visible light scenarios. Experimental results on newly annotated and existing datasets demonstrate that our model achieves state-of-the-art performance in intra- and cross-dataset scenarios while maintaining efficient execution.

show abstract

Segmentation of stroke lesions using transformers‐augmented MRI analysis

Ahmed,

Al Shehhi,

Werghi

et al. 2024

Human Brain Mapping

View full text Add to dashboard Cite

Accurate segmentation of chronic stroke lesions from mono‐spectral magnetic resonance imaging scans (e.g., T1‐weighted images) is a difficult task due to the arbitrary shape, complex texture, variable size and intensities, and varied locations of the lesions. Due to this inherent spatial heterogeneity, existing machine learning methods have shown moderate performance for chronic lesion delineation. In this study, we introduced: (1) a method that integrates transformers' deformable feature attention mechanism with convolutional deep learning architecture to improve the accuracy and generalizability of stroke lesion segmentation, and (2) an ecological data augmentation technique based on inserting real lesions into intact brain regions. Our combination of these two approaches resulted in a significant increase in segmentation performance, with a Dice index of 0.82 (±0.39), outperforming the existing methods trained and tested on the same Anatomical Tracings of Lesions After Stroke (ATLAS) 2022 dataset. Our method performed relatively well even for cases with small stroke lesions. We validated the robustness of our method through an ablation study and by testing it on new unseen brain scans from the Ischemic Stroke Lesion Segmentation (ISLES) 2015 dataset. Overall, our proposed approach of transformers with ecological data augmentation offers a robust way to delineate chronic stroke lesions with clinically relevant accuracy. Our method can be extended to other challenging tasks that require automated detection and segmentation of diverse brain abnormalities from clinical scans.

show abstract

SIPFormer: Segmentation of Multiocular Biometric Traits With Transformers

Cited by 5 publications

References 54 publications

An online color and shape integrated detection method for flexible packaging surface defects

An online color and shape integrated detection method for flexible packaging surface defects

OcularSeg: Accurate and Efficient Multi-Modal Ocular Segmentation in Non-Constrained Scenarios

Segmentation of stroke lesions using transformers‐augmented MRI analysis

Contact Info

Product

Resources

About