30 September 2022 in ACM Transactions on Multimedia Computing, Communications, and Applications
Detecting out-of-distribution (OOD) inputs for deep learning models is a critical task when models are deployed in real-world environments. Recently, a large number of works have been dedicated to tackling the OOD detection problem. One of the most straightforward and effective ways is OOD training, which adds heterogeneous auxiliary data in the training stage. However, the extra auxiliary data cannot be involved arbitrarily. A high-quality and powerful auxiliary dataset must contain samples that belong to OOD but are close to in-distribution (ID), which can teach the model to learn more information about OOD samples, furthermore, distinguish OOD from ID. The key issue for this problem is how to simply acquire such distinctive OOD samples. In this article, we propose an enhanced Mixup-based OOD (MixOOD) detection strategy that can be attached to any threshold-based OOD detecting method. Different from the traditional Mixup designed for ID data augmentation, our proposed MixOOD generates augmented images with deliberately modified Mixup and then uses them as auxiliary OOD data to leverage the OOD detection. We test our method with classical OOD detecting approaches like Maximum Softmax Probability, Energy Score, and Out-of-distribution detector for Neural networks. Experiments show that models with MixOOD can better distinguish in- and out-of-distribution samples than the original version of each approach.
1 September 2025 in IEEE Transactions on Circuits and Systems for Video Technology
Light Field (LF) is extensively utilized for depth estimation tasks due to its rich structural information. However, real-world LF images often encounter reflective and transparent surfaces and the related regions contain depth information from the reflection and background layers, which can be modeled as dual-layer scenes. For the existing depth estimation frameworks, the constructed cost volume shows an aliasing bimodal distribution in dual-layer surfaces and further causes serious wrong depth results. In this paper, we propose a novel decoupling-and-aggregating strategy and develop a dual-layer depth estimation network for LF images with complex reflections. Specifically, we develop an adaptive cost volume decoupling module to separate both the background and reflection features from the aliasing cost volume. Light field angular-spatial information is sufficiently extracted to infer the effort of features in different dimensions to the background or reflection layer. Additionally, we employ an iterative self-guided aggregating module with multi-stage supervision to aggregate two branches of cost volumes. The module applies the self-guided masks to regularize the distribution of cost volumes. Given the challenge of acquiring the ground truth disparity maps for the LF images under reflection scenes, we also construct a synthetic dataset with dual-layer properties. Our model is the first to introduce dual-layer scenes into the LF depth estimation task using an end-to-end deep neural network. It successfully separates the background and reflection layers and achieves accurate depth estimation results in both layers. Quantitative and qualitative experiment results on publicly available datasets demonstrate that our method performs better than other state-of-the-art methods.