Trademark watch services regularly search selected trademark databases to find potentially similar trademark applications. When evaluating entries for similarity, the following approaches are often used:
- Grammatical & syntax analysis, including misspellings, plurals, prefixes, abbreviations and more.
- Checking for translations
- Phonetic matching, where the algorithm can detect trademarks that sound similar, even if spelt differently.
-
- Class-based filtering, where classes of goods and services associated with the trademarks are taken into account.
- Image recognition techniques for evaluating visual similarity in figurative marks.