About
Yiming Wang is a researcher in the Deep Visual Learning (DVL) Unit in Fondazione Bruno Kessler (FBK) and the Pattern Analysis and Computer Vision/Visual Geometry Modelling (PAVIS/VGM) research line at Istituto Italiano di Tecnologia (IIT). Before that, Yiming obtained her PhD degree in 2018 from Queen Mary University of London (UK) under the supervision of Prof. Andrea Cavallaro. Her research enthusiasm is on vision-based scene understanding that facilitates robotic automation. Yiming is currently involved in a couple of EU projects that further advances her expertise towards human behaviour analysis with privacy-preservation in mind. Yiming also serves as reviewers for most top-tier conferences in Computer Vision and robotics, e.g. CVPR, ICCV, ECCV, IROS and ICRA, and journals including IJCV, TMM, RAL, TRO.
Papers (20)
2026
Specificity-aware reinforcement learning for fine-grained open-world classification
2025
ConViS-Bench: Estimating Video Similarity Through Semantic Concepts
LOTS of Fashion! Multi-Conditioning for Image Generation via Sketch-Text Pairing
On Large Multimodal Models as Open-World Image Classifiers
Training-Free Personalization via Retrieval and Reasoning on Fingerprints
Automatic benchmarking of large multimodal models via iterative experiment programming
Diversified in-domain synthesis with efficient fine-tuning for few-shot classification
Evaluating Attribute Confusion in Fashion Text-to-Image Generation
Seeing the abstract: Translating the abstract language for vision language models
One vlm to keep it learning: Generation and balancing for data-free continual visual question answering
Collaborative Neural Painting
2024
Retrieval-enriched zero-shot image classification in low-resource domains
Delving into CLIP latent space for Video Anomaly Recognition
Harnessing Large Language Models for Training-free Video Anomaly Detection
Test-Time Zero-Shot Temporal Action Localization
2023
Vocabulary-free Image Classification
ConfMix: Unsupervised Domain Adaptation for Object Detection via Confidence-based Mixing
2022
Graph-based generative face anonymisation with pose preservation
