Vmake Integrates ICML-Accepted BridgeRemoval Algorithm for Advanced AI Video Object Removal

PALO ALTO, Calif., May 20, 2026 /PRNewswire/ -- Vmake has officially integrated BridgeRemoval, a next-generation AI algorithm developed to improve the quality and speed of video object removal. The research paper behind this algorithm has been accepted at ICML 2026. The paper can be viewed on arXiv at the official research page. 

This integration highlights Vmake's focus on research-led development, where a globally recognized ICML paper becomes the foundation of its everyday video watermark removal feature.

The problem with video object removal today

Many existing video object removal systems still rely on diffusion methods that begin from random noise instead of the original footage itself. This often weakens scene accuracy, creates unrealistic background reconstruction, and causes flickering between frames. 

Since video depends on stable motion and visual continuity, these problems become noticeable very quickly, particularly in scenes with movement, texture detail, or changing lighting conditions.

How the BridgeRemoval Algorithm Improves Video Object Removal

The core innovation of BridgeRemoval lies in its unique methodology. Unlike traditional algorithms that attempt to reconstruct video frames from scratch, BridgeRemoval establishes a stochastic bridging path directly from the source video (containing the object) to the target video output.

This method brings two clear benefits. One is a steady motion across frames. The background holds its place as the video plays, which helps avoid the jitter or flicker often seen in other algorithms used for video object removal. 

The other is speed. The system handles processing quickly while still keeping results clean and realistic. That makes it useful for fast turnaround work where both quality and time matter. Creators can view the technical capabilities and visual results of this technology on the official project page.

Why ICML 2026 recognition matters

The 43rd International Conference on Machine Learning (ICML 2026) will take place in Seoul, South Korea, from July 6–11, 2026, at the COEX Convention & Exhibition Center. This is one of the world's biggest events for AI and Machine Learning research. Getting accepted by ICML means the "BridgeRemoval paper" passed a strict peer review from leading researchers around the world. It confirms the originality, technical strength, and practical value of the algorithm.

The recognition strengthens Vmake's position in AI video object removal technology research. Rather than simply adopting existing models, the company continues building tools based on next-gen technologies.

"BridgeRemoval's ICML acceptance means the technology behind Vmake's AI removal tool has been peer-reviewed," said Shiny, A Product Manager at Vmake. "Our goal is simple. Turn research-level ideas into practical tools for real work."

Vmake's Video Removal feature built on BridgeRemoval

Vmake's video watermark removal feature uses precise object targeting powered by BridgeRemoval to improve removal accuracy across both static and moving scenes. With it, users can easily remove watermarks, logos, subtitles, any text overlay, any unnecessary object, and even people from videos with pinpoint accuracy.

The feature supports multi-object and batch removal. This helps creators handle larger editing tasks faster. It also restores the background smoothly after removal, keeping the final result natural. The tool is built for easy use, so creators, sellers, and small teams can work with it without advanced editing skills.

ABOUT Vmake

Vmake is an all-in-one UGC video generator that helps small businesses and creators generate, enhance, and optimize high-performing videos for acquisition, conversion, and growth. It combines video generation, editing, enhancement, and optimization into one workflow.  This eliminates the need for separate apps for each process, helping users create stronger content with less time and lower production costs.

Connect with Us:

YouTube: https://www.youtube.com/@Vmake_ai

X (Twitter): https://x.com/VmakeAI

TikTok: https://www.tiktok.com/@Vmake.ai

LinkedIn: https://www.linkedin.com/company/Vmake-ai-captions/

Discord: https://discord.com/invite/TUbzh3mTtz

Instagram: https://www.instagram.com/Vmake.ai/

SOURCE Vmake