Cross-modal hashing with missing labels.

Neural Netw

School of Computer Science, Guangdong Polytechnic Normal University, 293 Zhonghshan Dadao, Tianhe District, Guangzhou, 510665, Guangdong, China. Electronic address:

Published: August 2023

Hashing-based cross-modal retrieval methods have become increasingly popular due to their advantages in storage and speed. While current methods have demonstrated impressive results, there are still several issues that have not been addressed. Specifically, many of these approaches assume that labels are perfectly assigned, despite the fact that in real-world scenarios, labels are often incomplete or partially missing. There are two reasons for this, as manual labeling can be a complex and time-consuming task, and annotators may only be interested in certain objects. As such, cross-modal retrieval with missing labels is a significant challenge that requires further attention. Moreover, the similarity between labels is frequently ignored, which is important for exploring the high-level semantics of labels. To address these limitations, we propose a novel method called Cross-Modal Hashing with Missing Labels (CMHML). Our method consists of several key components. First, we introduce Reliable Label Learning to preserve reliable information from the observed labels. Next, to infer the uncertain part of the predicted labels, we decompose the predicted labels into latent representations of labels and samples. The representation of samples is extracted from different modalities, which assists in inferring missing labels. We also propose Label Correlation Preservation to enhance the similarity between latent representations of labels. Hash codes are then learned from the representation of samples through Global Approximation Learning. We also construct a similarity matrix according to predicted labels and embed it into hash codes learning to explore the value of labels. Finally, we train linear classifiers to map original samples to a low-dimensional Hamming space. To evaluate the efficacy of CMHML, we conduct extensive experiments on four publicly available datasets. Our method is compared to other state-of-the-art methods, and the results demonstrate that our model performs competitively even when most labels are missing.

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.neunet.2023.05.035DOI Listing

Publication Analysis

Top Keywords

labels
16
missing labels
16
predicted labels
12
cross-modal hashing
8
hashing missing
8
cross-modal retrieval
8
latent representations
8
representations labels
8
representation samples
8
hash codes
8

Similar Publications

Heterogeneity in Fluorescence-Stained Sperm Membrane Patterns and Their Dynamic Changes Towards Fertilization in Mice.

Front Biosci (Landmark Ed)

January 2025

Graduate School of Life and Environmental Sciences, Integrated Graduate School of Medicine, Engineering, and Agricultural Sciences, University of Yamanashi, 400-8510 Kofu, Japan.

Background: Sperm represent a heterogeneous population crucial for male reproductive success. Additionally, sperm undergo dynamic changes during maturation and capacitation. Despite these well-established processes, the complex nature of sperm heterogeneity and membrane dynamics remains elusive.

View Article and Find Full Text PDF

Background: Myocardial ischemia-reperfusion (I/R) injury refers to cell damage that occurs as a consequence of the restoration of blood circulation following reperfusion therapy for cardiovascular diseases, and it is a primary cause of myocardial infarction. The search for nove therapeutic targets in the context of I/R injury is currently a highly active area of research. p70 ribosomal S6 kinase (S6K1) plays an important role in I/R induced necrosis, although the specific mechanisms remain unclear.

View Article and Find Full Text PDF

Introduction: The effects of remimazolam (Re) in combination with andrographolide (AP) on learning, memory, and motor abilities in rats following cardiopulmonary bypass (CPB) surgery were studied.

Methods: We hypothesized that the combination of Re and AP could improve postoperative cognitive dysfunction (POCD) in rats after CPB by modulating nervous system inflammation. Cognitive function was assessed using the Morris Water Maze test, and the concentrations of tumor necrosis factor alpha (TNF-α), interleukin-1 beta (IL-1β), and interleukin-6 (IL-6) in serum were measured by enzyme-linked immunosorbent assay (ELISA).

View Article and Find Full Text PDF

Radio frequency identification (RFID) technology and marker recognition algorithms can offer an efficient and non-intrusive means of tracking animal positions. As such, they have become important tools for invertebrate behavioral research. Both approaches require fixing a tag or marker to the study organism, and so it is useful to quantify the effects such procedures have on behavior before proceeding with further research.

View Article and Find Full Text PDF

Phosphodiesterase-4 (PDE4) is involved in the synthesis of inflammatory cytokines that mediate several chronic inflammatory disorders, including psoriasis and atopic dermatitis. In recent years, the therapeutic armamentarium in dermatology has expanded with the introduction of PDE4 inhibitors, both in oral and topical formulations. PDE4 inhibitors have gained increasing interest due to their remarkable safety record and ease of prescription, as evidenced by the recent influx of literature detailing its off-label uses.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!