- 偵測可能具冒犯性或不應出現的影像
- 篩選可能具冒犯性或不恰當的文字
- 管制影片中的限制級與種族歧視內容
- 能產生最佳結果的內建審核工具
影像審核
透過機器學習式的分類器、自訂封鎖清單及光學字元辨識 (OCR),加強偵測潛在冒犯性或垃圾影像的能力。
文字審核
使用內容篩選可偵測 100 多種語言中可能具有冒犯性的文字、根據內容將可能不恰當的用字加上標記 (公開預覽階段),並以您自訂的清單進行文字比對。內容仲裁也有助於檢查個人識別資訊 (PII)。
影片審核
利用機器協助,偵測影片中可能具限制級與種族歧視的內容。Azure 媒體服務提供影片仲裁服務 (公開預覽階段)。
人工審核工具
人機共同作業將可產生最佳的仲裁結果。當預測信任度在實際情況下有改進或調整的空間時,可利用這項審核工具。

""A safe and positive experience for our players in all interactions with the game is very important to who we are as a company.We utilize Content Moderator from Microsoft Cognitive Services for our player profile tools on Let's Dish, a new Match 3 cooking game. This is one way we are maintaining that trust and positive experience with our players.""
Big Fish Games 行銷經理 Ashley Davidson