Knowledge Mining with Scene Text for Fine-Grained Recognition

Liao, Junchao; Cheng, Tianheng; Zewen, Gao,; Líu, Hao; Ren, Bo; Bai, Xiang; Liu, Wenyu

doi:10.1109/cvpr52688.2022.00458

Cited by 9 publications

(4 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…developed a deep learning method based on an end‐to‐end trainable network that mines implicit contextual knowledge behind scene text image and enhance the semantics and correlation to fine‐tune the image representation, which outperformed the SOTA by 3.72% and 5.39%. [ 29 ] Q. Song et al.…”

Section: Discussionmentioning

confidence: 99%

“…[28] For example, H. Wang et al developed a deep learning method based on an end-to-end trainable network that mines implicit contextual knowledge behind scene text image and enhance the semantics and correlation to fine-tune the image representation, which outperformed the SOTA by 3.72% and 5.39%. [29] Q. Song et al proposed a deep learning method with multimodal sparse transformer network (MMST) and achieved a better performance (≈5% lower word error rate compared to SOTA) for different types of noise (−5 to 10 dB).…”

Section: Discussionmentioning

confidence: 99%

“…The spectrogram was obtained using the short-time Fourier transform where the Hamming window was performed. Then the average energy values of the five frequency bands (𝛿(0-4 Hz), 𝜃(4-8 Hz), 𝛼(8-12 Hz), 𝛽 (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), and 𝛾(30-40 Hz)) were calculated based on the spectrogram results. The power spectral densities of the five frequency bands were calculated according to the previous study.…”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Mimicking the Biological Sense of Taste In Vitro Using a Taste Organoids‐on‐a‐Chip System

Chen

Qin

et al. 2023

Advanced Science

View full text Add to dashboard Cite

Thanks to the gustatory system, humans can experience the flavors in foods and drinks while avoiding the intake of some harmful substances. Although great advances in the fields of biotechnology, microfluidics, and nanotechnologies have been made in recent years, this astonishing recognition system can hardly be replaced by any artificial sensors designed so far. Here, taste organoids are coupled with an extracellular potential sensor array to form a novel bioelectronic organoid and developed a taste organoids-on-a-chip system (TOS) for highly mimicking the biological sense of taste ex vivo with high stability and repeatability. The taste organoids maintain key taste receptors expression after the third passage and high cell viability during 7 days of on-chip culture. Most importantly, the TOS not only distinguishs sour, sweet, bitter, and salt stimuli with great specificity, but also recognizes varying concentrations of the stimuli through an analytical method based on the extraction of signal features and principal component analysis. It is hoped that this bioelectronic tongue can facilitate studies in food quality controls, disease modelling, and drug screening.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Mimicking the Biological Sense of Taste In Vitro Using a Taste Organoids‐on‐a‐Chip System

Chen

Qin

et al. 2023

Advanced Science

View full text Add to dashboard Cite

show abstract

“…Since their proposal, subsequent contribution have addressed limitations of their approach and adapted it to other challenges. Recent publications extend the ability of NeRFs to support dynamic scenes [PSB*21; PSH*21; PCPM21; TTG*21; LNSW21], accelerating inference time [RPLG21; MESK22; YLT*21; FYT*22; CXG*22; LSS*21; WZL*22; CFHT23], making them robust against the challenges of in‐the‐wild image capture [MRS*21; TCY*22; MHM*22; RLS*22], reducing the required image count [YYTK21; DLZR22; NBM*22; YPW23; RMY*22] and enabling dynamic relighting [ZSD*21; SDZ*21; MHS*22].…”

Section: Related Workmentioning

confidence: 99%

A Post Processing Technique to Automatically Remove Floater Artifacts in Neural Radiance Fields

Wirth,

Rak,

Knauthe

et al. 2023

Computer Graphics Forum

View full text Add to dashboard Cite

Neural Radiance Fields have revolutionized Novel View Synthesis by providing impressive levels of realism. However, in most in‐the‐wild scenes they suffer from floater artifacts that occur due to sparse input images or strong view‐dependent effects. We propose an approach that uses neighborhood based clustering and a consistency metric on NeRF models trained on different scene scales to identify regions that contain floater artifacts based on Instant‐NGPs multiscale occupancy grids. These occupancy grids contain the position of relevant optical densities in the scene. By pruning the regions that we identified as containing floater artifacts, they are omitted during the rendering process, leading to higher quality resulting images. Our approach has no negative runtime implications for the rendering process and does not require retraining of the underlying Multi Layer Perceptron. We show on a qualitative base, that our approach is suited to remove floater artifacts while preserving most of the scenes relevant geometry. Furthermore, we conduct a comparison to state‐of‐the‐art techniques on the Nerfbusters dataset, that was created with measuring the implications of floater artifacts in mind. This comparison shows, that our method outperforms currently available techniques. Our approach does not require additional user input, but can be be used in an interactive manner. In general, the presented approach is applicable to every architecture that uses an explicit representation of a scene's occupancy distribution to accelerate the rendering process.

show abstract

Visual crowd analysis: Open research problems

Khan,

Menouar,

Hamila

2023

AI Magazine

View full text Add to dashboard Cite

Over the last decade, there has been a remarkable surge in interest in automated crowd monitoring within the computer vision community. Modern deep‐learning approaches have made it possible to develop fully automated vision‐based crowd‐monitoring applications. However, despite the magnitude of the issue at hand, the significant technological advancements, and the consistent interest of the research community, there are still numerous challenges that need to be overcome. In this article, we delve into six major areas of visual crowd analysis, emphasizing the key developments in each of these areas. We outline the crucial unresolved issues that must be tackled in future works, in order to ensure that the field of automated crowd monitoring continues to progress and thrive. Several surveys related to this topic have been conducted in the past. Nonetheless, this article thoroughly examines and presents a more intuitive categorization of works, while also depicting the latest breakthroughs within the field, incorporating more recent studies carried out within the last few years in a concise manner. By carefully choosing prominent works with significant contributions in terms of novelty or performance gains, this paper presents a more comprehensive exposition of advancements in the current state‐of‐the‐art.

show abstract

Knowledge Mining with Scene Text for Fine-Grained Recognition

Cited by 9 publications

References 30 publications

Mimicking the Biological Sense of Taste In Vitro Using a Taste Organoids‐on‐a‐Chip System

Mimicking the Biological Sense of Taste In Vitro Using a Taste Organoids‐on‐a‐Chip System

A Post Processing Technique to Automatically Remove Floater Artifacts in Neural Radiance Fields

Visual crowd analysis: Open research problems

Contact Info

Product

Resources

About