Scrabble Vision

Client

Mattel

Relevant Tech Stack

Unity, OpenCV, Go, Treasure Data, Nakama

Work Summary

-Custom machine vision algorithms

-AI hint system

-Data analytics

-Scoring logic

-Leaderboards

-Commercial mobile gameplay for both iOS and Android

The Problem

Mattel has been passionately dedicated to innovating their customers’ experiences by modernizing the games that have brought joy to generations. In the case of Scrabble, Mattel was keen on helping new Scrabble players understand the rules and complexities of the game. They were also looking for a way to help bridge the gap between players that are just getting started, and hardened Scrabble veterans.

The Solution

Over the course of 7 months, HookBang developed a companion app called Scrabble® Vision: Scorekeeper+. The mobile application was built using custom machine vision algorithms trained to recognize multiple Scrabble boards, as well as their corresponding tile sets. The robust algorithms leverage any mobile phone’s camera system to detect correctly played words, and to then subsequently score them as well. This has dramatically improved the player experience by streamlining the traditionally cumbersome process of manual scoring.

An intelligent AI built into the app is also able to generate a list of hints for the player based on the tiles they have available on their player rack. By dynamically adjusting who receives how many hints, families have been able to enjoy more balanced games with players of all different ages.