NSTL国家科技图书文献中心

1. BRIDGE: Bridging Gaps in Image Captioning Evaluation with Stronger Visual Cues NSTL国家科技图书文献中心

Sara Sarto | Marcella Cornia ... - 《Computer Vision - ECCV 2024,Part LXXVIII》 - European Conference on Computer Vision - 2025, - 70～87 - 共18页

摘要：Effectively aligning with human judgment when evaluating machine-generated image captions represents a complex yet intriguing challenge. Existing evaluation metrics like CIDEr or CLIP-Score fall short...

关键词： Captioning evaluation | Vision-and-Language

2. Fluent and Accurate Image Captioning with a Self-trained Reward Model NSTL国家科技图书文献中心

Nicholas Moratelli | Marcella Cornia ... - 《Pattern Recognition,Part XVIII》 - International Conference on Pattern Recognition - 2025, - 209～225 - 共17页

摘要：Fine-tuning image captioning models with hand-crafted rewards like the CIDEr metric has been a classical strategy for promoting caption quality at the sequence level. This approach, however, is known ...

关键词： CLIP-based reward | Image captioning | Vision-and-Language models

3. Contrasting Deepfakes Diffusion via Contrastive Learning and Global-Local Similarities NSTL国家科技图书文献中心

Lorenzo Baraldi | Federico Cocchi ... - 《Computer Vision - ECCV 2024,Part LXIII》 - European Conference on Computer Vision - 2025, - 199～216 - 共18页

摘要：Discerning between authentic content and that generated by advanced AI methods has become increasingly challenging. While previous research primarily addresses the detection of fake faces, the identif...

关键词： Deepfake detection | Contrastive learning

4. Adapt to Scarcity: Few-Shot Deepfake Detection via Low-Rank Adaptation NSTL国家科技图书文献中心

Silvia Cappelletti | Lorenzo Baraldi ... - 《Pattern Recognition,Part XXI》 - International Conference on Pattern Recognition - 2025, - 111～126 - 共16页

摘要：The boundary between AI-generated images and real photographs is becoming increasingly narrow, thanks to the realism provided by contemporary generative models. Such technological progress necessitate...

关键词： Deepfake detection | Few-Shot learning | LoRA

5. Unlearning Vision Transformers Without Retaining Data via Low-Rank Decompositions NSTL国家科技图书文献中心

Samuele Poppi | Sara Sarto ... - 《Pattern Recognition,Part III》 - International Conference on Pattern Recognition - 2025, - 147～163 - 共17页

摘要：The implementation of data protection regulations such as the GDPR and the California Consumer Privacy Act has sparked a growing interest in removing sensitive information from pre-trained models with...

关键词： Machine unlearning | Low-Rank adaptation | Vision transformers | Image classification

6. Parents and Children: Distinguishing Multimodal Deepfakes from Natural Images NSTL国家科技图书文献中心

ROBERTO AMOROSO | DAVIDE MORELLI ... - 《ACM transactions on multimedia computing communications and applications》 - 2025,21(1) - 11.1～11.23 - 共23页 - 被引量：1

摘要：Recent advancements in diffusion models have enabled the generation of realistic deepfakes from textual prompts in natural language. While these models have numerous benefits across various sectors, t...

关键词： Multimodal deepfakes | vision-and-language | generative models

7. Safe-CLIP: Removing NSFW Concepts from Vision-and-Language Models NSTL国家科技图书文献中心

Samuele Poppi | Tobia Poppi ... - 《Computer Vision - ECCV 2024,Part LIII》 - European Conference on Computer Vision - 2025, - 340～356 - 共17页

摘要：Large-scale vision-and-language models, such as CLIP, are typically trained on web-scale data, which can introduce inappropriate content and lead to the development of unsafe and biased behavior. This...

关键词： Trustworthy AI | Vision-and-Language | NSFW concepts

8. FOSSIL: Free Open-Vocabulary Semantic Segmentation through Synthetic References Retrieval NSTL国家科技图书文献中心

Luca Barsellotti | Roberto Amoroso ... - 《2024 IEEE/CVF Winter Conference on Applications of Computer Vision: WACV 2024, Waikoloa, Hawaii, USA, 3-8 January 2024, [v.11]》 - IEEE/CVF Winter Conference on Applications of Computer Vision - 2024, - 1453～1462 - 共10页

摘要：Unsupervised Open-Vocabulary Semantic Segmentation aims to segment an image into regions referring to an arbitrary set of concepts described by text, without relying on dense annotations that are avai...

关键词： Training | Visualization | Sensitivity | Semantic segmentation | Semantics | Prototypes | Predictive models

9. AIGeN: An Adversarial Approach for Instruction Generation in VLN NSTL国家科技图书文献中心

Niyati Rawal | Roberto Bigazzi ... - 《2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops》 - IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops - 2024, - 2070～2080 - 共11页

摘要：In the last few years, the research interest in Vision-and-Language Navigation (VLN) has grown significantly. VLN is a challenging task that involves an agent following human instructions and navigati...

关键词： Measurement | Navigation | Computational modeling | Training data | Bidirectional control | Transformers | Encoding

10. Video Surveillance and Privacy: A Solvable Paradox? NSTL国家科技图书文献中心

Rita Cucchiara | Lorenzo Baraldi ... - 《Computer》 - 2024,57(3) - 91～100 - 共10页 - 被引量：1

摘要：Through experiments on action recognition and natural language description, we show that the paradox of surveillance and privacy can be solved by artificial intelligence and that respect for human rig...

摘要：Effectively aligning with human judgment when evaluating machine-generated image captions represents a complex yet intriguing challenge. Existing evaluation metrics like CIDEr or CLIP-Score fall short...

关键词： Captioning evaluation | Vision-and-Language

摘要：Fine-tuning image captioning models with hand-crafted rewards like the CIDEr metric has been a classical strategy for promoting caption quality at the sequence level. This approach, however, is known ...

关键词： CLIP-based reward | Image captioning | Vision-and-Language models

摘要：Discerning between authentic content and that generated by advanced AI methods has become increasingly challenging. While previous research primarily addresses the detection of fake faces, the identif...

关键词： Deepfake detection | Contrastive learning

摘要：The boundary between AI-generated images and real photographs is becoming increasingly narrow, thanks to the realism provided by contemporary generative models. Such technological progress necessitate...

关键词： Deepfake detection | Few-Shot learning | LoRA

摘要：The implementation of data protection regulations such as the GDPR and the California Consumer Privacy Act has sparked a growing interest in removing sensitive information from pre-trained models with...

关键词： Machine unlearning | Low-Rank adaptation | Vision transformers | Image classification

摘要：Recent advancements in diffusion models have enabled the generation of realistic deepfakes from textual prompts in natural language. While these models have numerous benefits across various sectors, t...

关键词： Multimodal deepfakes | vision-and-language | generative models

摘要：Large-scale vision-and-language models, such as CLIP, are typically trained on web-scale data, which can introduce inappropriate content and lead to the development of unsafe and biased behavior. This...

关键词： Trustworthy AI | Vision-and-Language | NSFW concepts

摘要：Unsupervised Open-Vocabulary Semantic Segmentation aims to segment an image into regions referring to an arbitrary set of concepts described by text, without relying on dense annotations that are avai...

关键词： Training | Visualization | Sensitivity | Semantic segmentation | Semantics | Prototypes | Predictive models

摘要：In the last few years, the research interest in Vision-and-Language Navigation (VLN) has grown significantly. VLN is a challenging task that involves an agent following human instructions and navigati...

关键词： Measurement | Navigation | Computational modeling | Training data | Bidirectional control | Transformers | Encoding

摘要：Through experiments on action recognition and natural language description, we show that the paradox of surveillance and privacy can be solved by artificial intelligence and that respect for human rig...

关键词： Privacy | Natural languages | Video surveillance | Artificial intelligence

4008-161-200 800-990-8900

国家科技图书文献中心