Performance analysis of U-Net with hybrid loss for foreground detection

Abstract

With the latest developments in deep neural networks, the convolutional neural network (CNN) has made considerable progress in the area of foreground detection. However, the top-rank background subtraction algorithms for foreground detection still have many shortcomings. It is challenging to extract the true foreground against complex background. To tackle the bottleneck, we propose a hybrid loss-assisted U-Net framework for foreground detection. A proposed deep learning model integrates transfer learning and hybrid loss for better feature representation and faster model convergence. The core idea is to incorporate reference background image and change detection mask in the learning network. Furthermore, we empirically investigate the potential of hybrid loss over single loss function. The advantages of two significant loss functions are combined to tackle the class imbalance problem in foreground detection. The proposed technique demonstrates its effectiveness on standard datasets and performs better than the top-rank methods in challenging environment. Moreover, experiments on unseen videos also confirm the efficacy of proposed method.

An approach to the dermatological classification of histopathological skin images using a hybridized CNN-DenseNet model

“…Although this work focuses on a different medical imaging domain, it demonstrates the potential of multitask learning models for complex diagnostic tasks, which can also be applied to dermatological classification. Kalsotra & Arora (2023) conducted a performance analysis of the U-Net model with hybrid loss for foreground detection. While their study was not specific to dermatology, their findings highlighted the importance of loss function selection in segmentation tasks.…”

“…The fully connected layers aggregate these features and make the final predictions. Natural CNNs can automatically learn meaningful hierarchical representations from raw image data that enable them to capture complex patterns and variations, leading to highly accurate and robust classification results ( Kalsotra & Arora, 2023 ; Ke et al, 2021 ).…”