I work between doctoral research and industry technology leadership, combining computer-vision research with production experience in backend architecture, enterprise software, AI products, business operations, and commercialisation-facing systems.
My recent work spans interpretable image representations, multimodal scene understanding, diffusion-based methods, and dataset-building for visual learning.
Learning interpretable representations and robust visual understanding from images, scenes, and multimodal observations.
Studying how visual, spatial, audio, textual, and viewpoint-specific signals can be fused for richer scene understanding.
Exploring controllable generative systems that connect representation learning with editing, restoration, and creative workflows.
Applying machine learning methods in scientifically grounded settings where interpretability and structure matter.
Chenyuan Qu, Hao Chen, Jianbo Jiao
VisualSplit learns image representations from decoupled edges, colour segmentation, and grey-level histograms, enabling descriptor-to-image reconstruction, controllable editing, and diffusion-based restoration.
2025
British Machine Vision Conference (BMVC)
Hao Chen, Yuqi Hou, Chenyuan Qu, Irene Testini, Xiaohan Hong, Jianbo Jiao
A panoptic multimodal dataset that combines panoramic, frontal, and egocentric viewpoints with audio, location, and textual signals for richer scene understanding benchmarks.
2024
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
2025
VisualSplit
Chenyuan Qu, Hao Chen, Jianbo Jiao
British Machine Vision Conference (BMVC)
VisualSplit learns image representations from decoupled edges, colour segmentation, and grey-level histograms, enabling descriptor-to-image reconstruction, controllable editing, and diffusion-based restoration.
2025
Yuxiang Ji, Boyong He, Chenyuan Qu, Zhuoyue Tan, Chuan Qin, Liaoni Wu
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
DIFF leverages diffusion-model features to improve cross-domain semantic segmentation by extracting and fusing semantically rich representations across the diffusion process.
2024
360+x
Hao Chen, Yuqi Hou, Chenyuan Qu, Irene Testini, Xiaohan Hong, Jianbo Jiao
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
A panoptic multimodal dataset that combines panoramic, frontal, and egocentric viewpoints with audio, location, and textual signals for richer scene understanding benchmarks.
2023
MeD
Hao Chen, Chenyuan Qu, Yu Zhang, Chen Chen, Jianbo Jiao
IEEE/CVF International Conference on Computer Vision (ICCV)
A self-supervised denoising framework that disentangles clean image structure from corruption by comparing multiple noisy views of the same latent scene.
A multimodal scene-understanding dataset spanning panoramic, frontal, and egocentric video, with aligned audio, location, and textual signals for benchmarking perception across viewpoints.
A binocular egocentric and 360° panoramic multimodal dataset and challenge surface for scene understanding, aligned with spatial audio, text, and geo-metadata.
A privacy-safe text-to-image dataset released on Hugging Face, repacked into Parquet shards with embedded image bytes and organised into samples and iterations splits.
I started the 12-week Help To Grow: Management Course at Birmingham City University Business School, with sessions spanning strategy, digital transformation, marketing, operations, finance, and growth planning.
Project page, paper, supplementary material, code, and model weights are publicly available.
Collaborative work on diffusion features for cross-domain semantic segmentation.
Dataset paper with accompanying benchmark resources, code, and public dataset access.
I began doctoral research in 2023 on computer vision and multimodal learning in the MI X Group.
An early project on self-supervised image denoising using multi-view disentanglement.
Doctoral research in computer vision, multimodal learning, and generative AI within the School of Computer Science.
Research on compositionality for foundation models, with a focus on representation, reasoning, and generative computer-vision systems.
Research on interpretable hydrological modelling and machine-learning methods for scientific analysis.
Own the company-wide technology roadmap for digital transformation and AI acceleration, spanning backend architecture, enterprise software, business operations, sales enablement, and commercialisation-facing systems.
Led major full-stack and traditional software-engineering work for the CMS and ERP platforms, including the CMS rebuild that moved the company away from fragmented board-specific codebases toward one shared software architecture across hardware boards and operating systems.
Built early backend and AI capabilities for the Allsee and Vieunite product stack, developing recommendation infrastructure and Vieutopia AI art features that connected machine-learning experiments with production product needs.
Developed and optimised machine-learning components for a visual customer-service anchor in a mobile deployment setting.
Postgraduate study completed at Birmingham before beginning doctoral research.
Undergraduate training in physics, which continues to shape how I think about machine learning and computer vision.
The easiest way to reach me is by email.