Image moderation is the process of filtering images that are inappropriate for a brand’s website or social media accounts. This includes images that depict nudity, violence, or any other explicit themes that are not aligned with a brand’s core objectives.
Cloudinary’s image moderation add-on, WebPurify, enables you to automatically detect and filter unwanted content in photos, videos and live streams. Its API returns moderation results instantly and scales to fit your needs.
Manual Moderation
In the online world of social media, message boards, forums, streaming services, and other user-generated content, reputation is everything. If your business hosts content that is inappropriate, illegal, upsetting, or harmful, it can quickly tarnish your brand and cause damage to your customers’ experience and the trust they have in your service. To avoid these risks, you need to keep content that could be considered trolling, spamming, flaming, or even predatory behaviour off of your platform. Moderation tools powered by AI, such as Pure Moderation’s image content moderation software Imagga, are a viable solution to help you monitor images, videos and live streams. This tech-powered method of moderation combines image recognition, text analysis, and even computer vision technology to spot sensitive imagery and flag it for manual review.
Automated Moderation
Many online platforms employ automated moderation tools to speed up the process of sifting through user-generated content. They work by detecting and flagging inappropriate images, videos, or text on their own, without being explicitly instructed to do so by human moderators.
The problem with this type of automation is that it doesn’t have any cultural or contextual understanding. For instance, it’s hard for AI to recognize how words or slang might be offensive in one context but not in another, as well as the nuanced variations in cultural norms of specific regions.
It can also be difficult to detect when an image or video has been manipulated. This is why ML algorithms need to be taught by a vast library of rules and examples, so they can spot similar content at scale. The automated systems used by Imagga, for example, use a combination of ML techniques to automatically identify and flag content on the basis of its own guidelines. This includes computer vision, which identifies objects or characteristics in images, such as nudity or weapons. It also uses OCR to make text in images machine-readable for NLP processing.
Rejection
Whether they’re avatars, comments, photos, or videos, images leave a strong impact on users. They’re also a critical part of your brand’s reputation and can boost your search engine rankings.
When it comes to images, moderation is crucial to maintain the quality of your community and protect your reputation. It helps you avoid legal disputes, increase user retention and improve overall site performance.
For instance, a kid-friendly social network uses image moderation to monitor content for instances of gore, nudity, weapons, drug use, or violence. When an inappropriate image is detected, it gets flagged for human moderation, deleted automatically, or displayed with a warning to users.
With Zia’s image moderation tool, you can set up rules based on the type of images you want to screen. You can also block specific categories of images, such as adult or racy content, or put uncertain images on hold to be reviewed manually. This is especially useful if you have limited moderation resources.