Awarded the Best Paper Award and Best Poster Presentation Award at the Workshop on Image Processing and Understanding (IPIU 2025) / Master's Program in Computer Science, Graduate School of Engineering
Master's Program in Computer Science, Graduate School of Engineering Junghyun Seo, Soye Kwon (24), Seungheon Song (23), Euihyun Yoon (22)
- 25.03.20 / 이정민
The Artificial Intelligence Research Laboratory (supervised by Professor Jae-koo Lee) of the Department of Computer Science and Engineering at Kookmin University's Graduate School of General Studies achieved excellent results, winning four awards at the 37th Workshop on Image Processing and Understanding (IPIU 2025).
Seo Jeong-hyun (24), a student in the master's program, won the 'Outstanding Thesis Award', while Kwon So-ye (24), Song Seung-heon (23), and Yoon Eui-hyun (22) won the 'Outstanding Poster Presentation Award'.
Seo Junghyun conducted research on “Robust camera-based 3D occupancy estimation for environmental changes through noise perturbation.” This research proposed a noise perturbation method to create a robust model that can withstand various environmental changes that occur in autonomous driving perception situations in three-dimensional space.
Kwon So-ye conducted a study on “Image Editing Based on Diffusion Model Using Multimodal Prompts.” This study proposed a new method that allows users to edit customized images using multimodal control. Unlike the existing method, where the user's intention was limited to a single or dual modality, this study expanded to various modalities to allow more sophisticated image editing. Users can control structural information through sketches in the area they want to edit, define content information through text prompts, and transfer style features through image prompts.
Song Seung-heon conducted a study on 'Improving the performance of minority-shot classification using CLIP-ViT decomposition'. This study used CLIP, a multimodal model, to improve the performance of classification with a small amount of data.
Yoon Eui-hyun conducted a study on 'Improving the performance of single image object segmentation using a visual-linguistic model.' This study found that the existing object segmentation task failed to properly generate masks and solved this problem by using the linguistic characteristics of the mask to improve performance.
|
This content is translated from Korean to English using the AI translation service DeepL and may contain translation errors such as jargon/pronouns. If you find any, please send your feedback to kookminpr@kookmin.ac.kr so we can correct them.
|
|
Awarded the Best Paper Award and Best Poster Presentation Award at the Workshop on Image Processing and Understanding (IPIU 2025) / Master's Program in Computer Science, Graduate School of Engineering Master's Program in Computer Science, Graduate School of Engineering Junghyun Seo, Soye Kwon (24), Seungheon Song (23), Euihyun Yoon (22) |
||||
|---|---|---|---|---|
|
The Artificial Intelligence Research Laboratory (supervised by Professor Jae-koo Lee) of the Department of Computer Science and Engineering at Kookmin University's Graduate School of General Studies achieved excellent results, winning four awards at the 37th Workshop on Image Processing and Understanding (IPIU 2025).
Seo Jeong-hyun (24), a student in the master's program, won the 'Outstanding Thesis Award', while Kwon So-ye (24), Song Seung-heon (23), and Yoon Eui-hyun (22) won the 'Outstanding Poster Presentation Award'.
Seo Junghyun conducted research on “Robust camera-based 3D occupancy estimation for environmental changes through noise perturbation.” This research proposed a noise perturbation method to create a robust model that can withstand various environmental changes that occur in autonomous driving perception situations in three-dimensional space.
Kwon So-ye conducted a study on “Image Editing Based on Diffusion Model Using Multimodal Prompts.” This study proposed a new method that allows users to edit customized images using multimodal control. Unlike the existing method, where the user's intention was limited to a single or dual modality, this study expanded to various modalities to allow more sophisticated image editing. Users can control structural information through sketches in the area they want to edit, define content information through text prompts, and transfer style features through image prompts.
Song Seung-heon conducted a study on 'Improving the performance of minority-shot classification using CLIP-ViT decomposition'. This study used CLIP, a multimodal model, to improve the performance of classification with a small amount of data.
Yoon Eui-hyun conducted a study on 'Improving the performance of single image object segmentation using a visual-linguistic model.' This study found that the existing object segmentation task failed to properly generate masks and solved this problem by using the linguistic characteristics of the mask to improve performance.
|
||||
