Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks216 from public data
- 1,320
- Contrastive Adaptation Network for Unsupervised Domain Adaptation1,013
This paper proposes Contrastive Adaptation Network optimizing a new metric which explicitly models the intra- class domain discrepancy and the inter-class domain discrepancy, and designs an alternating update strategy for training CAN in an end-to-end manner.
- Evaluating bag-of-visual-words representations in scene classification962
This study provides an empirical basis for designing visual-word representations that are likely to produce superior classification performance and applies techniques used in text categorization to generate image representations that differ in the dimension, selection, and weighting of visual words.
- Cross-domain video concept detection using adaptive svms749
This paper proposes Adaptive Support Vector Machines (A-SVMs) as a general method to adapt one or more existing classifiers of any type to the new dataset and outperforms several baseline and competing methods in terms of classification accuracy and efficiency in cross-domain concept detection in the TRECVID corpus.
- 480
- A discriminative CNN video representation for event detection464
This paper proposes using a set of latent concept descriptors as the frame descriptor, which enriches visual information while keeping it computationally affordable, in a new state-of-the-art performance in event detection over the largest video datasets.
- DecideNet: Counting Varying Density Crowds Through Attention Guided Detection and Density Estimation416
A novel end-to-end crowd counting framework, named DecideNet (DEteCtIon and Density Estimation Network) is proposed, which can adaptively decide the appropriate counting mode for different locations on the image based on its real density conditions.
- MAGVIT: Masked Generative Video Transformer413
A 3D tokenizer to quantize a video into spatial-temporal visual tokens and propose an embedding method for masked video token modeling to facilitate multi-task learning are introduced.
- Self-Paced Learning with Diversity405
This work proposes an approach called self-paced learning with diversity (SPLD) which formalizes the preference for both easy and diverse samples into a general regularization term, independent of the learning objective, and thus can be easily generalized into various learning tasks.
- The Eighth Visual Object Tracking VOT2020 Challenge Results350
A significant novelty is introduction of a new VOT short-term tracking evaluation methodology, and introduction of segmentation ground truth in the VOT-ST2020 challenge – bounding boxes will no longer be used in theVDT challenges.
- 337
- 329
- Hidden Two-Stream Convolutional Networks for Action Recognition316
This paper presents a novel CNN architecture that implicitly captures motion information between adjacent frames and directly predicts action classes without explicitly computing optical flow, and significantly outperforms the previous best real-time approaches.
- Beyond Gaussian Pyramid: Multi-skip Feature Stacking for action recognition301
This work proposes a novel feature enhancing technique called Multi-skIp Feature Stacking (MIFS), which stacks features extracted using a family of differential filters parameterized with multiple time skips and encodes shift-invariance into the frequency space and proves that MIFS enhances the learnability of differential-based features exponentially.
- KAT: A Knowledge Augmented Transformer for Vision-and-Language232
This work proposes a - Knowledge Augmented Transformer (KAT) - which achieves a strong state-of-the-art result on the open-domain multimodal task of OK-VQA and integrates implicit and explicit knowledge in an encoder-decoder architecture, while still jointly reasoning over both knowledge sources during answer generation.
- 231
- 220
- 204
- 198
- Uncovering the Temporal Context for Video Question Answering188
An encoder–decoder approach using Recurrent Neural Networks to learn the temporal structures of videos and introduce a dual-channel ranking loss to answer multiple-choice questions is presented.
- ExCL: Extractive Clip Localization Using Natural Language Descriptions186
This work proposes a novel extractive approach that predicts the start and end frames by leveraging cross-modal interactions between the text and video - this removes the need to retrieve and re-rank multiple proposal segments.
- 171
- CADP: A Novel Dataset for CCTV Traffic Camera based Accident Analysis166
A novel dataset for traffic accidents analysis is presented to resolve the lack of public data for research about automatic spatio-temporal annotations for traffic safety in the roads and to integrate contextual information into conventional Faster R-CNN using Context Mining and Augmented Context Mining to complement the accuracy for small pedestrian detection.
- 157
- Learning Spatial Awareness to Improve Crowd Counting151
A novel architecture called SPatial Awareness Network (SPANet) to incorporate spatial context for crowd counting is presented, and the Maximum Excess over Pixels (MEP) loss is proposed to achieve this by finding the pixel-level subregion with high discrepancy to the ground truth.
- High level knowledge sources in usable speech recognition systems151
The authors detail an integrated system which combines natural language processing with speech understanding in the context of a problem solving dialogue.
- Rethinking Spatial Invariance of Convolutional Networks for Object Counting140
Inspired by previous work, this work proposes a low-rank approximation accompanied with translation invariance to favorably implement the approximation of massive Gaussian convolution to improve the spatial invariance of convolutional networks.
- 132
- Extreme video retrieval129
An efficient system for video search that maximizes the use of human bandwidth, while at the same time exploiting the machine's ability to learn in real-time from user selected relevant video clips is presented.
- Multimedia classification and event detection using double fusion118
This paper introduces a fusion scheme named double fusion, which simply combines early fusion and late fusion together to incorporate their advantages, and reports the best reported results to date.
- Semantic Concept Discovery for Large-Scale Zero-Shot Event Detection117
This work proposes an efficient, highly scalable algorithm that is an order of magnitude faster than existing alternatives for detecting complex events in unconstrained Internet videos in a more difficult zero-shot setting.
- 116
- Deep Local Video Feature for Action Recognition114
Experimental results on the HMDB51 and UCF101 datasets show that a simple maximum pooling on the sparsely sampled local features leads to significant performance improvement.
- Describing Videos using Multi-modal Fusion108
Experimental results show the effectiveness of multi-modal fusion encoder trained in the end-to-end framework, which achieved top performance in both common metrics evaluation and human evaluation.
- SimAug: Learning Robust Representations from Simulation for Trajectory Prediction104
A novel approach to learn robust representation through augmenting the simulation training data such that the representation can better generalize to unseen real-world test data.
- Exploring Semantic Inter-Class Relationships (SIR) for Zero-Shot Action Recognition101
SIR is proposed to perform action recognition when no positive exemplars of that class are provided, which is often known as the zero-shot learning, which directly leverages the semantic inter-class relationships between the known and unknown actions followed by label transfer learning.
- InfAR dataset: Infrared action recognition at different times100
This work constructs a new InfAR dataset captured at different times, and explores how discriminable actions in this dataset are with the state-of-the-art pipelines based on low-level features and deep convolutional neural network (CNN), respectively.
- Improving the Learning of Multi-column Convolutional Neural Network for Crowd Counting97
Extensive experiments on four challenging benchmarks show that McML can significantly improve the original multi-column networks and outperform the other state-of-the-art approaches.
- A review of text and image retrieval approaches for broadcast news video90
This paper reviews a variety of text/image retrieval approaches as well as their individual components in the context of broadcast news video, and conducts a series of retrieval experiments on TRECVID video collections to identify their advantages and disadvantages.
- Multiple instance learning for labeling faces in broadcasting news video88
This paper proposes two discriminative probabilistic learning methods named Exclusive Density (ED) and Iterative ED for eMIL problems and shows that the performance of the proposed approaches are superior to the traditional MIL algorithms and close to the performance achieved by supervised methods trained with complete data labels.
- Concepts Not Alone: Exploring Pairwise Relationships for Zero-Shot Video Activity Recognition87
This paper proposes a more robust approach for scoring concepts in order to alleviate many of the brittleness and low precision problems of previous work, and proposes a novel pairwise order matrix approach for score aggregation.
- 84
- 81
- Viral Video Style: A Closer Look at Viral Videos on YouTube80
The proposed method is unique in that it is the first attempt to incorporate video metadata into the peak day prediction, and outperforms the state-of-the-art methods, with statistically significant differences.
- 79
- Survey of current speech technology79
Speech recognition and speech synthesis are technologies of particular interest for their support of direct communication between humans and computers through a communications mode humans commonly use among themselves and at which they are highly skilled.
- Video Captioning with Guidance of Multimodal Latent Topics77
An unified caption framework, M&M TGM, which mines multimodal topics in unsupervised fashion from data and guides the caption decoder with these topics, which are more semantically and visually coherent and can reflect the topic distribution of videos better.
- 73
- 71
- Harnessing Lab Knowledge for Real-World Action Recognition69
A multi-task learning framework to jointly optimize the classifiers for both sides and is able to mine the shared knowledge between two datasets even if the two have different action categories, which is a major virtue of the method.
- Deep feature learning via structured graph Laplacian embedding for person re-identification68
This work proposes a structured graph Laplacian embedding algorithm, which can formulate all these structured distance relationships among these training samples into the graph LaPLacian form, and achieves state-of-the-art performances.
- Comparing Evaluation Protocols on the KTH Dataset67
How different test configurations can skew results, even on standard data set is shown, and the recommendation is to use a simple leave-one-out as the most easily replicable clear-cut partitioning.
- Massive Open Online Proctor66
This paper proposes a Massive Open Online Proctoring (MOOP) framework, which combines both automatic and collaborative approaches to detect cheating behaviors in online tests and shows that ACD and PCD can detect usage of a cheat sheet with good accuracy and can reduce the overall human resources required to monitor MOOCs for cheating.
- 65
- 64
- Instructional Videos for Unsupervised Harvesting and Learning of Action Examples62
This work proposes to utilize the large amount of instructional videos available online to harvest examples of various actions in an unsupervised fashion to exploit the timing of action corresponding terms in the speech transcript to temporally localize actions in the video and harvest action examples.
- Video retrieval using speech and image information58
An evaluation on the effects of different types of information used for video retrieval from a video collection found image matching and video OCR proved to be the deciding aspects of video information retrieval.
- 55
- 55
- 54
- 53
- 53
- 52
- Talking to computers: an empirical investigation50
The results show that speech to a computer is not as ill-formed as one would expect, and people speaking to acomputer are more disciplined than when speaking to each other.
- 49
- Few-shot activity recognition with cross-modal memory network49
This paper proposes a novel end-to-end cross-modal memory network for few-shot activity recognition that stores the dynamic visual and textual semantics for some high-level attributes related to human activities and could achieve significant improvements over other previous methods.
- 48
- 47
- 47
- 45
- Novelty detection for cross-lingual news stories with visual duplicates and speech transcripts44
The novelty and redundancy detection with visual duplicates and speech transcripts for cross-lingual news stories, and approaches on textual features and visual features demonstrate different performance, and measures on visual features are quite effective.
- Learning to recognize speech by watching television42
This work describes its approach to collecting almost unlimited amounts of accurately transcribed speech data, which serves as training data for the acoustic model component of most high-accuracy speaker-independent speech-recognition systems.
- Dictionary pruning with visual word significance for medical image retrieval40
This work proposes a retrieval method based on a bag-of-visual-words (BoVW) to identify discriminative characteristics between different medical images with Pruned Dictionary based on Latent Semantic Topic description.
- Probabilistic Non-Negative Matrix Factorization and Its Robust Extensions for Topic Modeling37
This paper follows the generative procedure of topic model and learns the topic-word distribution and topics distribution via directly approximating the word-document co-occurrence matrix with matrix decomposition technique and adopts the explicit probabilistic meaning of factors in topic models.
- 37
- 36
- Scheduled sampling for one-shot learning via matching network36
A novel metric is proposed to evaluate the learning difficulty of support set, where this metric jointly considers the semantic diversity and similarity of visual labels and introduces a scheduled sampling strategy to train the matching network from easy to difficult.
- Efficient human action recognition using histograms of motion gradients and VLAD with descriptor shape information36
This work proposes Shape Difference VLAD (SD-VLAD), an encoding method which brings complementary information by using the shape information within the encoding process, and validated the proposed pipeline for action recognition on three challenging datasets UCF50, UCF101 and HMDB51.
- Space-Time Robust Representation for Action Recognition36
This work introduces an iterative structure learning algorithm, WSVM (weighted SVM), that determines the optimal saliency layout of an action model through a sparse regularizer and proposes a novel content driven pooling that leverages space-time context while being robust toward global space-time transformations.
- 35
- ZSTAD: Zero-Shot Temporal Activity Detection34
This work designs an end-to-end deep network based on R-C3D that is optimized with an innovative loss function that considers the embeddings of activity labels and their super-classes while learning the common semantics of seen and unseen activities.
- 34
- 31
- 29
- 28
- 28
- 26
- 26
- 26
- 26
- 25
- 25
- 23
- 22
- 22
- Avoiding Optimal Mean ℓ2,1-Norm Maximization-Based Robust PCA for Reconstruction21
This letter equivalently reformulate the objective of conventional PCA and learn the optimal projection directions by maximizing the sum of projected difference between each pair of instances based on -norm by exploiting an efficient optimization algorithm to soften the contributions from outliers by reweighting each data point iteratively.
- 19
- 19
- 19
- 18
- 18
- 18
- 18
- 17
- 16
- 16
- 16
- 15
- 15
- Cross-domain video concept detection: A joint discriminative and generative active learning approach15
- 15
- 15
- Adaptive context-aware reinforced agent for handwritten text recognition14
This paper addresses handwritten text recognition problems by leveraging an adaptive context-aware reinforced agent which learns the actions to determine the scales of context regions during inference by formulate the approach in a reinforcement learning framework.
- 14
- 14
- 13
- 13
- 13
- 13
- 12
- 12
- 12
- 12
- 11
- 11
- UMETTS: A Unified Framework for Emotional Text-to-Speech Synthesis with Multimodal Prompts10
The UMETTS framework is introduced, a novel framework that leverages emotional cues from multiple modalities to generate highly expressive and emotionally resonant speech and achieves significant improvements in emotion accuracy and speech naturalness, outperforming traditional E-TTS methods on both objective and subjective metrics.
- 10
- 10
- 10
- SZTU-CMU at MER2024: Improving Emotion-LLaMA with Conv-Attention for Multimodal Emotion Recognition9
This system leverages the advanced emotional understanding capabilities of Emotion-LLaMA to generate high-quality annotations for unlabeled samples, addressing the challenge of limited labeled data, and introduces Conv-Attention, a lightweight and efficient hybrid framework to enhance multimodal fusion while mitigating modality-specific noise.
- 9
- MetaDesigner: Advancing Artistic Typography Through AI-Driven, User-Centric, and Multilingual WordArt Synthesis9
MetaDesigner introduces a transformative framework for artistic typography synthesis, powered by Large Language Models (LLMs) and grounded in a user-centric design paradigm, that dynamically adjusts hyperparameters to align with user-defined stylistic and thematic preferences.
- 9
- The Best of BothWorlds: Combining Data-Independent and Data-Driven Approaches for Action Recognition9
- 9
- 9
- 9
- 8
- 8
- 8
- 8
- 8
- 7
- 7
- 7
- 6
- 6
- 6
- 6
- 6
- 6
- 5
- 5
- 5
- 5
- 5
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-13. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.