| 1. |
-
- 2019
-
Tidskriftsartikel (refereegranskat)
|
|
| 2. |
- Wang, Guan'an, et al.
(författare)
-
Euclidean-Distance-Preserved Feature Reduction for efficient person re-identification
- 2024
-
Ingår i: Neural Networks. - Oxford : Elsevier. - 0893-6080 .- 1879-2782. ; 180
-
Tidskriftsartikel (refereegranskat)abstract
- Person Re-identification (Re-ID) aims to match person images across non-overlapping cameras. The existing approaches formulate this task as fine-grained representation learning with deep neural networks, which involves extracting image features using a deep convolutional network, followed by mapping the features into a discriminative space through another smaller network, in order to make full use of all possible cues. However, recent Re-ID methods that strive to capture every cue and make the space more discriminative have resulted in longer features, ranging from 1024 to 14336, leading to higher time (distance computation) and space (feature storage) complexities. There are two potential solutions: reduction-after-training methods (such as Principal Component Analysis and Linear Discriminant Analysis) and reduction-during-training methods (such as 1 × 1 Convolution). The former utilizes a statistical approach aiming for a global optimum but lacking end-to-end optimization of large data and deep neural networks. The latter lacks theoretical guarantees and may be vulnerable to training noise such as dataset noise or initialization seed. To address these limitations, we propose a method called Euclidean-Distance-Preserving Feature Reduction (EDPFR) that combines the strengths of both reduction-after-training and reduction-during-training methods. EDPFR first formulates the feature reduction process as a matrix decomposition and derives a condition to preserve the Euclidean distance between features, thus ensuring accuracy in theory. Furthermore, the method integrates the matrix decomposition process into a deep neural network to enable end-to-end optimization and batch training, while maintaining the theoretical guarantee. The result of the EDPFR is a reduction of the feature dimensions from fa and fb to fa′ and fb′, while preserving their Euclidean distance, i.e. L2 (fa, fb) = L2 (fa′, fb′). In addition to its Euclidean-Distance-Preserving capability, EDPFR also features a novel feature-level distillation loss. One of the main challenges in knowledge distillation is dimension mismatch. While previous distillation losses, usually project the mismatched features to matched class-level, spatial-level, or similarity-level spaces, this can result in a loss of information and decrease the flexibility and efficiency of distillation. Our proposed feature-level distillation leverages the benefits of the Euclidean-Distance-Preserving property and performs distillation directly in the feature space, resulting in a more flexible and efficient approach. Extensive on three Re-ID datasets, Market-1501, DukeMTMC-reID and MSMT demonstrate the effectiveness of our proposed Euclidean-Distance-Preserving Feature Reduction. © 2024 Elsevier Ltd
|
|
| 3. |
- Wang, Kun, et al.
(författare)
-
Distributed Energy Management for Vehicle-to-Grid Networks
- 2017
-
Ingår i: IEEE Network. - Piscataway, NJ : IEEE. - 0890-8044 .- 1558-156X. ; 31:2, s. 22-28
-
Tidskriftsartikel (refereegranskat)abstract
- Making full use of V2G services, EVs with batteries may assist the smart grid in alleviating peaks of energy consumption. Aiming to develop a systematic understanding of the interplay between smart grid and EVs, an architecture for the V2G networks with the EV aggregator is designed to maintain the balance between energy suppliers (the grid side) and consumers (the EV side). We propose a combined control and communication approach considering distributed features and vehicle preferences in order to ensure efficient energy transfer. In our model, the integrated communication and control unit can achieve realtime and intelligent management with the logic controller and collected data. On the consumers' side, we theoretically analyze how to satisfy the charging constraints that we incorporate in the form of willingness to pay, and propose a distributed framework to coordinate the energy delivery behaviors for satisfying service demands. Moreover, illustrative results indicate that the proposed approach can yield higher revenue than the conventional pricing mechanism in V2G networks.
|
|
| 4. |
- Xu, Minjie, et al.
(författare)
-
Platelets derived citrullinated proteins and microparticles are potential autoantibodies ACPA targets in RA patients
- 2023
-
Ingår i: Frontiers in Immunology. - Lausanne : Frontiers Media S.A.. - 1664-3224. ; 14
-
Tidskriftsartikel (refereegranskat)abstract
- Citrullinated neoepitopes have emerged as key triggers of autoantibodies anti-citrullinated protein antibodies (ACPA) synthesis in rheumatoid arthritis (RA) patients. Apart from their critical role in homeostasis and thrombosis, platelets have a significant contribution to inflammation as well. Although anuclear in nature, platelets have an intricate post-translational modification machinery. Till now, citrullination in platelets and its contribution to trigger autoantibodies ACPA production in RA is an unexplored research direction. Herein, we investigated the expression of peptidylarginine deiminase (PAD) enzymes and citrullinated proteins/peptides in the human platelets and platelet derived microparticles (PDP). Both PAD4 mRNA and protein, but not the other PAD isoforms, are detectable in the human platelets. With a strict filtering criterion,108 citrullination sites present on 76 proteins were identified in the human platelets, and 55 citrullinated modifications present on 37 different proteins were detected in the PDPs. Among them, some are well-known citrullinated autoantigens associated with RA. Citrullinated forms of thrombospondin-1, β-actin, and platelet factor-4 (also known as CXCL4) are highly immunogenic and bound by autoantibodies ACPA. Furthermore, ACPA from RA sera and synovial fluids recognized citrullinated proteins from platelets and significantly activated them as evidenced by P-selectin upregulation and sCD40 L secretion. These results clearly demonstrate the presence of citrullinated autoantigens in platelets and PDPs, thus could serve as potential targets of ACPA in RA. Copyright © 2023 Xu, Du, Xing, Chen, Wan, Wang, Xiong, Nandakumar, Holmdahl and Geng.
|
|
