AI vision technology enables machines to perceive and understand the visual world the way humans see it. It combines computer vision and AI techniques to detect and recognize visual elements and analyze attributes such as color, shape, movement and context within images and videos.
By leveraging Microsoft solutions such as: Azure Cloud and Azure OpenAI service, California-based Chooch provides AI vision capabilities for a wide range of applications in various industries, enabling machines to accurately interpret and understand visual data. The recently released Imagechat injects Large Language Models (LLMs) with AI vision, which clients use to address their forensic, training, and analytics needs across live and stored visual content. Connect to image and video data lakes.
Emrah Gultekin, co-founder and CEO of Chooch, talks about the incredible amount of visual data we face every day, how AI can help us make sense of it, and how advances in computer and AI vision can help other We talked about what startups can learn.
Take advantage of the explosive growth of visual data
Emra doesn’t choke on the technical conundrum that Chuchi is grappling with.
“The problem is the explosion of cameras and visual data in the world today,” Emra told me. “Even if everyone on the planet reviewed this data, there wouldn’t be enough people to do it. It’s about automating the detection and recognition of events in content.”
“This is no longer just about one AI, it is about speech, language, transcription, translation, tabular data and computer vision. I have.”
To achieve this, Chooch integrates large-scale generative AI vision models and fuses them with LLM to enable new inference and more accurate contextual understanding for edge and cloud hosted applications.
“While our work on computer vision AI has been primarily about building software infrastructure, our main innovation has been to deploy lightweight inference engines in self-hosted and edge environments, making traditional computer vision models and It was the ability to fuse LLMs together,” explains Emrah. “The same explosion we’re seeing in terms of language is happening in computer vision, and he’s the complex problem we’re solving that fuses the two.”
Entrepreneurs can find endless avenues to harness computer vision in today’s increasingly scrutinized world. Emra points to the power of this technology to help security and safety personnel analyze images and data from public spaces, workplaces, airports and industrial sites to detect and respond to threats. Industries such as manufacturing and distribution are leveraging computer vision to improve efficiency and reduce human error. The Chooch AI platform improves the accuracy and speed of visual processes such as defect analysis and quality control, ensuring a safer work environment.
Build AI products responsibly
To build successful AI vision solutions, Emrah advises other startups that the visual and linguistic aspects of AI go hand in hand. The two fields are closely related as they both rely on the ability to extract meaning from data. Visual AI systems trying to extract meaning from visuals within a scene or sequence of frames need to understand the context of object names and descriptions. Similarly, a linguistic AI system trying to understand sentences needs to understand the meaning of the words in the sentence and the relationships between them.
“Without language, vision is not as influential,” says Emra. “My advice to startups is to experiment with the multimodal side of AI, because we have that capability now. Because they traditionally don’t speak the same language, but this is no longer just about one AI, it’s about speech, language, transcription, translation, tabular data, computer vision. , the impact we have on our clients is so great that we all need to be united.”
Partnering with Microsoft to focus on building the best solutions
Before embarking on the new AI era, Chooch had to overcome some of the problems of traditional AI startups, including a lack of both initial infrastructure and tech stack. In addition to building many stacks, Emra said, it took an iterative, trial-and-error approach to reason and analyze progress in this uncharted territory.
Emrah says the partnership with Microsoft is very important. Because Microsoft leads the industry in computing power. Chooch uses Azure Machine Learning, Azure Cognitive Services, Azure IoT Hub and Edge to ingest data from edge devices.
“We are essentially aligned in terms of strengthening the AI marketplace and AI for Good,” says Emrah. “Compared to Microsoft’s competitors, we received a lot of support for what we were building. We were able to leverage our resources.”
A member of the Microsoft for Startups Pegasus program since late 2022, he is grateful that Microsoft is giving companies the flexibility to focus on developing top-of-the-line solutions that benefit the entire partner ecosystem. says there is.
“Microsoft’s CTO, Kevin Scott, said it perfectly,” recalls Emrah. “‘Don’t worry about the infrastructure, just build a good product.'”
Microsoft for Startups Founders Hub members receive Azure Cloud Credits that can be used for Azure OpenAI Service or OpenAI to help build products. Sign up now to become a member.