• News

    Finnish Startup Binit Recognizes Trash Using OpenAI’s Image Recognition

    Rayan

    By Rayan - Jun 3, 2024 | Updated On: 09 June, 2024 | 2 min read

    By Rayan , 2 min read - Jun 3, 2024

    Updated On: 09 June, 2024

    Finnish Startup Binit Recognizes Trash. Image Credit: Social Media.




  • With the ever-evolving AI, innovative ideas are not only limited to robots, computers, and drones. AI has spread in all areas, and companies are noticing it.

    Finnish startup Binit recognizes trash using the Large Langage Models (LLMs) algorithm. It sure is an intelligent way to dispose of waste from individual homes!

    Finnish Startup Binit Recognizes Trash

    Binit is a startup in Finland like Enifer but is involved in applying large language models’ (LLMs) image processing capabilities to track household trash. Yes, you heard it right!




  • Founder Borut Grgic believes this area is yet to be tapped and is an unexploited territory.

    Although Binit is a startup like Zipline and was founded during the recent pandemic, Binit has amassed a whopping $3 million in funding from an angel investor in the early phase.

    The company is building AI hardware explicitly designed to be located in the kitchen, mounted to cabinets or near walls where bin-related activities occur.

    How Does This Bin-Related AI Operate?

    The AI-incorporated hardware has onboard cameras and other sensors designed to activate when someone is nearby, enabling them to scan before throwing it in the bin.




  • Founder Grgic said the module integrates with commercial LLMs, mainly OpenAI’s GPT, for image recognition.

    Finnish Startup Binit Recognizes Trash

    Finnish Startup Binit Recognizes Trash. Image Credit: Linkedin.

    Following image recognition, Finnish Startup Binit Recognizes Trash, which generates an entire system of analytical feedback and gamification through an app.

    Weekly rubbish scores, just like our weekly calorie tracking scores, are done to stimulate users to reduce the unnecessary amount of trash tossed out.

    ALSO READ: OpenAI GPT-4o Is The New Model: Announces ChatGPT Desktop Versions




  • How Did The Testing And Prototype Go?

    Binit tried to train their own AI models for trash recognition, but the results weren’t satisfactory, with accuracy being just 40%.

    So, to manage this inaccurate data, the team decided to use OpenAI’s image recognition capabilities.

    After the integration of LLM, trash recognition accuracy surged to an impressive 98%, as claimed by Grgic.

    What Does The Founder Have To Say?

    Binit founder Grgic explains that his startup is the first household waste tracker, opening the gates for building a sleep tracker like AI but for trash-tossing habits.




  • LLMs will be exploited to identify wastage since this trash tracking is a camera vision technology backed by the neural network.

    The founder even expresses his astonishment at how well the Open AI recognized trash with such high scores.

    Grgic says,” It’s not clear whether lots of images of trash were in OpenAI’s training data or the technology is just too fantastic at recognition.

    As Finnish startup Binit recognizes trash, we wish the startup miles of success and acceptance by the general public as a useful device for their trash habits.










  • Did you notice an error ?

    Please help us make corrections by submitting a suggestion. Your help is greatly appreciated!