セブリング革命、AIでポストプロダクション作業の時間とコストを削減

課題

The Sebring post-production team was losing hours of time removing sightings of other cameras or studio doors from their clips. It was tedious work. Additionally, segmenting people from the background for Augmented Reality was cumbersome and an extra cost.

The team decided to outsource this work, but this added between 5—7 days to their production process and cost them between $0.20 — $3.50 per frame.

The Sebring Revolution team needed a way to cut costs, minimize delivery time, and own their creation process in-house for a smoother client and employee experience.

ゴール

To improve and streamline the post-production process for all stakeholders. To decrease the time to delivery and post-production costs.

解決策

The Skim Ai team built an object detection AI model. This enabled the Sebring Revolution team to automate identifying parts of images, cameras, and doors that need to be addressed. This model automatically processes each image so that the small camera slots and doors are made invisible.

スキムのAIチームは、AIとMLのモデルも構築した：

• In-painting background
• Image Segmentation (subject from background)
• White Balancing
• Image Stabilization

Overall, this offered the team a more streamlined post-production experience, which no longer needed to be sourced out of-house. This enabled Sebring Revolution to own the entire editing process, speeding up their delivery by more than a week and drastically cutting costs on outsourcing.