IndiaTravel and aviation2019 – ongoingGoibibo
Automated moderation and tagging of hotel and review photos
Goibibo cut hotel photo publishing from a few hours to 30 minutes with image-recognition models
Problem
Goibibo moderated and categorized photos from hotels and reviewers with a slow, manpower-intensive system. Hotels and users waited a few hours for their photos to go live.
Approach
Goibibo added Google Cloud's Vision API and AutoML to its content pipeline. Vision API moderates images, flags offensive content, and assigns labels. AutoML trained a custom model on past data to sort photos into categories such as swimming pools, restaurants, and washrooms.
Results
Takeaway
Photo moderation was a repetitive task that a trained classifier could take over. Hotels and reviewers saw the gain as faster publishing.
Categories
- Size
- Large enterprise
- Industry
- Travel and aviation
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Google Cloud's customer story describes how Goibibo used Cloud Vision API and Cloud AutoML to moderate, annotate, and categorize hotel and user photos. Hotel photos went live in 30 minutes on average and user reviews with photos in 10 minutes, each down from a few hours. The page carries no date; the Internet Archive first saved it on July 11, 2019.
Full details are in the original sources.
