Anymal

WHO

WHAT

HOW

Inquisitive nature enthusiasts

Redesigning current methods of animal identification

By utilising an intelligent guessing system

JUMP 

CONTEXT

Anymal is a concept of mine that was conceived from a problem that I encountered while on a walk. I spotted an animal and realised that trying to identify the animal felt much harder than it should. Googling just wasn't producing accurate results and before I knew it, the animal escaped.

PROBLEM

Identifying a specific animal can be a long and tedious process. Current applications that aim to solve the problem make use of image detection technology.However, this may not be sufficient and have a tendency to be inaccurate and dependent on environmental conditions.

RESEARCH

SEEK - An animal identification app that uses image detection.

A few rounds of tests led to inconsistent results, the primary factor being the quality of the photo and the subject. The results themselves have the tendency to produce vague outputs that offer no insight as to the identity of the animal.

Difficulty capturing "good" photos:
Animals escape when they perceive danger. Taking a clear and close-up photo may not be a viable option.

Multiple methods of identification :
As identification is the critical feature required for the success of the app, secondary methods might be required.

Precise identification:
The result from the application should be more precise than visual identification by the user.

Speed is Key :
The process to identify animals should be optimised to be efficient yet accurate as it is a critical feature of the app.

THEORY

Anymal - An intelligent, community driven app

By leveraging on tried and tested methods and adapting it to fit the problem, I hope to improve the success rate and overall experience of identifying animals via an application.

QnA - Inspired by Akinator

By asking a series of "Yes" or "No" questions, we can narrow down the pool of potential animals and eventually reach a conclusion. Since animals have distinctive traits that are well-documented, we can leverage on these research to build a database of questions.

This method does not require a high quality photo, rather a photo that is useful enough for the user to answer the questions given. This also allows users the ability to take numerous photos first and lowers the risk of animals escaping while identifying the animal.

Leveraging on the community

If the primary method fails, the user can send their photo to the community to seek help in identifying the animal in the photo. The type of question ranges from open ended to polls depending on the various stages the query is at. This secondary method works as a "fail-safe" approach to give users the result they require.

THOUGHTS

By adapting Akinator's s QnA method and applying it to the identification of animals, it might prove to be a superior solution than existing methods. Image detection technology can even be merged with Anymal, potentially leading to greater success.

Finally, by further leveraging on collective intelligence, we ensure that the user is able to receive a possible result.