6 Companies Making Big Profits with Image Analysis AI (feat. Computer Vision AI)
AI Insights
5 min read

6 Companies Making Big Profits with Image Analysis AI (feat. Computer Vision AI)

We've all heard countless times that AI can look at an image and tell a dog from a cat. But we got curious about how it's used in business. We've put together a list of companies making big profits with image analysis AI features.

This article is recommended for the following people.

  • Those who've wondered how to make use of image analysis AI

  • Those who were curious about how image analysis AI is actually being used

Here's how they were using image analysis AI.

  • It retouched photos taken by ordinary people to look like they were shot in a studio.

  • It checked whether a product was defective just by looking at the image.

  • It even logged calories simply from a photo of food.

6 companies that use image analysis AI

  1. Cal AI

  2. Bepro11

  3. Socar

  4. Photoroom

  5. Landing AI

  6. Alarmy

Let's take a look at them one by one.

Cal AI: Food recognition & calorie counting app


Service: An app where you upload a photo of your food and the AI calculates the calories. You can record what kinds of food you ate and how much. It's growing so fast that just six months after launch it reached monthly revenue of 900 million won.

이 이미지는 칼로리가 계산되어 내 음식 소비 현황을 기록할 수 있는 걸 보여주는 Cal AI 앱에 대한 이미지입니다.
Calories are calculated so you can track your food consumption.

Image analysis AI in use: When you take a photo, your phone's depth sensor figures out the amount of food. An AI model trained on thousands of food images distinguishes the food's parts and proportions. A multimodal AI model then aggregates all the information and produces the calories, protein, carbs, and fat of the meal.

Limitations: Because calories are calculated from a photo, there's a limitation that foods like smoothies or soup are inaccurately measured. In those cases, the user has to enter the information directly in the app.

창업자는 Open AI에서 GPT-4 Vision API가 공개되었다는 사실을 듣고 Cal AI를 기획하게 되었다는 걸 나타내는 이미지입니다.
The founder heard that OpenAI had released the GPT-4 Vision API and went on to plan Cal AI. It appears they still use this API to identify food photos.
클로드에서도 음식 사진을 올리고 칼로리 계산을 요청하면 해주는 걸 나타내는 이미지입니다
Claude will also count calories if you upload a food photo and ask. Cal AI is a service that dug deep into this sharp, focused use case.

Bepro11: Person recognition & sports match analysis


Service: A startup that helps with sports match analysis. It started with soccer.

Image analysis AI in use: It watches the recorded match footage and recognizes the players and the ball in real time. It analyzes the match in one go.

Bepro11 카메라로 인식된 선수들 이미지입니다
Players recognized by the Bepro11 camera.
분석을 원하는 팀에게 제공하는 레포트에 대한 이미지입니다
It provides reports to teams that want analysis.
영상 속 사람의 골절을 인식하는 것이 bepro11의 원리라는 걸 나타내는 이미지이다.
Recognizing people's joints in the footage is the principle behind Bepro11.

Socar: Recognizing vehicle damage and dirt


Image analysis AI in use: It looked at captured images and automatically determined whether dents or scratches had occurred. It also figured out whether a vehicle was dirty enough to need a wash. Socar was using AI image analysis technology to streamline its operations.

딥러닝 모델을 통해 파손 여부를 분간해낸 걸 나타내는 이미지이다.
A deep learning model distinguished whether there was damage.
AI가 사진을 보고 어디에 이상이 있는지 체크해 주는 걸 나타내는 이미지이다.
The AI looking at a photo and checking where the problem is.
차량 파손에서 더 나아가 세차가 필요한 상황인지 파악하는 모델을 나타내는 이미지이다.
Going beyond vehicle damage, they also built a model that determines whether a wash is needed.

Photoroom: Removing & retouching backgrounds of online shop product photos


Service: A photo editing app for online shop operators who have to take countless product photos. It's growing fast—reaching 70 billion won in annual revenue just four years after founding.

Image analysis AI in use: The core feature is that the AI makes it easy to change a product's background. It doesn't stop at removing the background. You can choose the background you want. It even adds shadows, so it feels like the photo was originally shot against that background.

집에서 상품 사진을 찍으면, 스튜디오에서 만든 것처럼 정리해주는 이미지이다.
Take a product photo at home, and it polishes it up to look like it was made in a studio.

Landing AI: Checking factory defects (manufacturing)


Service: An AI company founded by Andrew Ng, a giant in the AI industry. It provides a platform for building image analysis models.

Image analysis AI in use: Let's say there's a manufacturing company that makes metal parts. It's important not to deliver defective products to customers. They automated, using AI, what people used to check one part at a time by eye. They installed cameras in the factory to photograph every part and detect whether any part has a defect.

새로 생산되는 부품들을 찍고 플랫폼에 올려놓기만 하면 되는 걸 나타내는 이미지이다.
All you have to do is photograph the newly produced parts and upload them to the platform.
부품에 이상이 있다면, 그 이상이 어디인지 위치를 알려주는 이미지이다.
If there's a defect in a part, it tells you the location.

Alarmy: Object recognition


Service: An alarm app where the alarm won't turn off until you complete a wake-up mission. Last year it was used by so many people that it achieved 24 billion won in annual revenue and 13.1 billion won in operating profit.

Image analysis AI in use: There's a mission where the alarm turns off when you photograph the same object every morning. The AI helps recognize it as the same object even at angles that inevitably vary slightly.

매일 아침 같은 물체를 찍어야 알람이 꺼지는 미션을 나타내는 이미지이다.
A mission where you have to photograph the same object every morning to turn off the alarm. AI is used in the process of confirming it's the same cup.

At Dalpha, where AI experts gather, we offer free AI consulting.


We'll consult and build a solution that's a perfect fit for your company.

Leave us your questions.

We'll get back to you within 10 minutes!

Apply for a free AI consultation

Articles worth reading together ?

kyungsuk chon

kyungsuk chon

You might also like...

How can we help?

We'll get back to you shortly.