Lin Li

Compositional Zero-shot Learning via Progressive Language-based Observations

Lin Li, Guikun Chen, Zhen Wang, Jun Xiao et autres

Compositional zero-shot learning aims to recognize unseen stateobject compositions by leveraging known primitives (state and object) during training. However, effectively modeling interactions between primitives and generalizing knowledge to novel compositions remains a perennial challenge. There are two crucial factors: large object-conditioned and …

hk, cn (code pays fourni par la source)

Zero-shot Compositional Action Recognition with Neural Logic Constraints

G Ye, Lin Li, Kexin Li, Jun Xiao et autres

Zero-shot compositional action recognition (ZS-CAR) aims to identify unseen verb-object compositions in the videos by exploiting the learned knowledge of verb and object primitives during training. Despite compositional learning's progress in ZS-CAR, two critical challenges persist: 1) Missing compositional structure constraint, leading …

cn, hk (code pays fourni par la source)

UFOS-Net leverages small-scale feature fusion for diabetic foot ulcer segmentation

Chenxu Jiao, Xiaodong Zhao, Lin Li, Chiyu Wang et autres

Diabetic Foot Ulcer (DFU) is a critical risk factor for disability and mortality among diabetic patients, posing a significant public health challenge. Existing DFU datasets are limited in capturing the diversity and complexity of ulcer manifestations, preventing advancements in medical segmentation. This …

cn (code pays fourni par la source)

Zero-shot Compositional Action Recognition with Neural Logic Constraints

G Ye, Lin Li, Kexin Li, Jun Xiao et autres

Zero-shot compositional action recognition (ZS-CAR) aims to identify unseen verb-object compositions in the videos by exploiting the learned knowledge of verb and object primitives during training. Despite compositional learning's progress in ZS-CAR, two critical challenges persist: 1) Missing compositional structure constraint, leading …

Knowledge Integration for Grounded Situation Recognition

Jiaming Lei, Sijing Wu, Lin Li, Lei Chen et autres

cn, hk (code pays fourni par la source)

Seeing Beyond Classes: Zero-Shot Grounded Situation Recognition via Language Explainer

Jiaming Lei, Lin Li, Chunping Wang, Jun Xiao et autres

Benefiting from strong generalization ability, pre-trained vision-language models (VLMs), e.g., CLIP, have been widely utilized in zero-shot scene understanding. Unlike simple recognition tasks, grounded situation recognition (GSR) requires the model not only to classify salient activity (verb) in the image, but also …

cn, hk (code pays fourni par la source)

From Easy to Hard: Learning Curricular Shape-Aware Features for Robust Panoptic Scene Graph Generation

Hanrong Shi, Lin Li, Jun Xiao, Yueting Zhuang et autres

cn, hk (code pays fourni par la source)

E-MFNN: an emotion-multimodal fusion neural network framework for emotion recognition

Zhuen Guo, Mingqing Yang, Lin Li, Jisong Li et autres

Emotional recognition is a pivotal research domain in computer and cognitive science. Recent advancements have led to various emotion recognition methods, leveraging data from diverse sources like speech, facial expressions, electroencephalogram (EEG), electrocardiogram, and eye tracking (ET). This article introduces a novel …

cn, us (code pays fourni par la source)

NICEST: Noisy Label Correction and Training for Robust Scene Graph Generation

Lin Li, Jun Xiao, Hanrong Shi, Hanwang Zhang et autres

Nearly all existing scene graph generation (SGG) models have overlooked the ground-truth annotation qualities of mainstream SGG datasets, i.e., they assume: 1) all the manually annotated positive samples are equally correct; 2) all the un-annotated negative samples are absolutely background. In this …

cn, sg, hk (code pays fourni par la source)

Label Semantic Knowledge Distillation for Unbiased Scene Graph Generation

Lin Li, Jun Xiao, Hanrong Shi, Wenxiao Wang et autres

The Scene Graph Generation (SGG) task aims to detect all the objects and their pairwise visual relationships in a given image. Although SGG has achieved remarkable progress over the last few years, almost all existing SGG models follow the same training paradigm: …

cn, hk (code pays fourni par la source)

Question-guided feature pyramid network for medical visual question answering

Yonglin Yu, Haifeng Li, Hanrong Shi, Lin Li et autres

cn (code pays fourni par la source)

The Devil is in the Labels: Noisy Label Correction for Robust Scene Graph Generation

Lin Li, Long Chen, Yifeng Huang, Zhimeng Zhang et autres

Unbiased SGG has achieved significant progress over recent years. However, almost all existing SGG models have overlooked the ground-truth annotation qualities of prevailing SGG datasets, i.e., they always assume: 1) all the manually annotated positive samples are equally correct; 2) all the …