# Point2Insert: Video Object Insertion via Sparse Point Guidance

Source: [Point2Insert: Video Object Insertion via Sparse Point Guidance](<https://www.sota2.com/research/paper/69842466a8b01e76cc8ee941>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-06T05:00:04.455Z

Authors: Yu Zhou, Xiaoyan Yang, Bojia Zi, Lihan Zhang, Ruijie Sun, Weishi Zheng, Haibin Huang, Chi Zhang, Xuelong Li

Published: 2026-02-04T03:01:15.000Z

## Abstract

This paper introduces Point2Insert, a sparse-point-based framework for flexible and user-friendly object insertion in videos, motivated by the growing popularity of accurate, low-effort object placement. Existing approaches face two major challenges: mask-based insertion methods require labor-intensive mask annotations, while instruction-based methods struggle to place objects at precise locations. Point2Insert addresses these issues by requiring only a small number of sparse points instead of dense masks, eliminating the need for tedious mask drawing. Specifically, it supports both positive and negative points to indicate regions that are suitable or unsuitable for insertion, enabling fine-grained spatial control over object locations. The training of Point2Insert consists of two stages. In Stage 1, we train an insertion model that generates objects in given regions conditioned on either sparse-point prompts or a binary mask. In Stage 2, we further train the model on paired videos synthesized by an object removal model, adapting it to video insertion. Moreover, motivated by the higher insertion success rate of mask-guided editing, we leverage a mask-guided insertion model as a teacher to distill reliable insertion behavior into the point-guided model. Extensive experiments demonstrate that Point2Insert consistently outperforms strong baselines and even surpasses models with $\\times$10 more parameters.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Generation Quality on PointBench](<https://www.sota2.com/research/sota/generation-quality-on-pointbench.md>): Generation Quality; PointBench; Success Rate (%): 94.44.

- [Mask-based video object insertion on Internal (test)](<https://www.sota2.com/research/sota/mask-based-video-object-insertion-on-internal-test.md>): Mask-based video object insertion; Internal (test); MSE: 103.09.

- [Point-based video object insertion on DAVIS (test)](<https://www.sota2.com/research/sota/point-based-video-object-insertion-on-davis-test.md>): Point-based video object insertion; DAVIS (test); Acc Pos: 81.13.

- [Video Editing on User Study (test)](<https://www.sota2.com/research/sota/video-editing-on-user-study-test.md>): Video Editing; User Study (test); Generative Quality: 3.858.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2602.04167>)

[PDF](<https://arxiv.org/pdf/2602.04167.pdf>)
