Photo by Akshay Kumawat on Unsplash

Re-imagining AI Cameras on mobile

Problem Statement

Gone are the days when people used to own bulky DSLR cameras and used to roam around with tripods and lighting devices. The age of AI cameras is here and it is here to stay. Instagram is one of the most downloaded applications in the world and many current-day mobile phones sell purely on the basis of their cameras. Manufacturers like Xiaomi have recently come up with phones fitted with a 108-megapixel sensor and have smashed all previous records on DXOMark.

Google’s AI camera has given out many path-breaking features like its night mode, burst shot etc. There are a multitude of applications in the market for quick visual content creation; TikTok being a prime one. The common link that connects all of these technologies or applications is a camera, which is scaling new heights every day thanks to advancements in AI.

The challenge was to come up with a new feature set for AI cameras and to develop these features in a way that they form a formidable business case. This case study was carried out in a period of two days.

Market study

To understand what the market already offers, we scoured the latest mobile phones with highest scores on DXOMark. We then looked at all that their cameras have to offer and made a list of all the features. This gave us a good understanding of what exists in the market and what does not.

User Research

Despite the glamorous offerings of most modern mobile cameras, it is important to understand how, when and why users use their mobile phone cameras first. Mobile phones have become more accessible in a country like India in the last couple of years. There are specific applications and services aimed at the breadth of user base this ginormously populated country has to offer.

After a quick check into our contact lists, we made a list of 10 people across different age groups, geographical locations and technical capabilities. We wanted to understand how, why and when they used their cameras.

The User Base

The following were the profiles of the people we interviewed:

1. 26-year-old software engineer. Photography enthusiast. Usually, the person who takes photos of everyone during a trip with a DSLR. Doesn’t post much.
2. 27-year-old IT support person. Lowkey professional photographer. Covers events like neighbourhood birthday parties and marriage ceremonies with his DSLR. Regularly posts images captured using his phone on social media.
3. 51-year-old entrepreneur/business consultant/life coach. Rarely takes or posts photos. Has a low-end phone. Averse to selfies.
4. 26-year-old lawyer. Regularly posts images captured using her phone on social media.
5. 55-year-old school teacher. Likes to take photos of her children and dogs at most. Has a low-end phone.
6. A 64-year-old father of two. Retired, living with wife, son, daughter in law and grandson. Extroverted and has an extended family. Actively shares content.
7. 59-year-old housewife. Lives with husband, mother in law and daughter. Heavily extroverted. Takes active interest with many social groups.
8. 23-year-old millenial. Doing a culinary course. Active socially on Instagram.
9. 28-year-old Millenial. In a tier 1 city. Very active with social media. Loves capturing moments and goes the extra mile.
10. 43-year-old farmer working in a tier 3 city. Lives with his wife and son lives far away in a tier 1 city.

User Interview prompts

We started by asking users about the last 5 times they opened a camera on their phone; irrespective of the mobile application through which it was opened. We dug deeper into each of these use cases and tried to find patterns in their usage across the ten users we interviewed.

Apart from that, we asked users to look into their galleries and dug behind the context of each of the last ten photos. For users who permitted us, we went through the different folders of the gallery and dug deep on the peculiar kinds of images that were captured or recorded. We also tried to understand the applications through which users used their camera or their photos. The stories kept coming and so did the insights.

Here are some of the sample questions we went ahead with:

1. Which phone do you use?
2. Can you tell us about the last three times you opened your camera? Did you click anything? If yes, can you tell me what it is?
3. On your gallery, if you could pick up the last 5 photos, can you let us know when, where and why you clicked them?
4. How many photos do you click every day?
5. Do you edit your photos in any way? If yes, how often and why?
6. What are the features you use most on your phone’s camera or any application that uses a camera or images?
7. Can you remember the last 5 images you have deleted? Can you remember why you deleted them?

After gathering the insights from all the users, we affinity mapped the insights and prioritised them based on their frequency.

Insights

1. Almost all the users performed some level of editing on their photos. Either through the gallery/camera application or through a third-party application.
2. Cropping an image was the most used edit feature. This was done mostly to remove unwanted elements from the frame or to position the subject of the photo better in the frame.
3. Levels was the second most used feature and almost all the users use it to some level. In levels, brightness was the most used feature and was used to lighten the skin tone of the subject. Saturation was the most used feature when shooting no human subjects.
4. Users also used the sharpness feature as a measure to undo unintended blur in hazy photos.
5. People deleted their photos mainly due to the following reasons: Blurred images, low lighting, dull light on subject, bad angle, composition etc.
6. Even if people could not vocalise what a good composition and image is, they were subconsciously aware of what constituted a good image and a bad one.
7. Editing was mostly restricted to photos. People barely ever edited their videos. Some people trimmed the video length at best, but hardly anyone used advanced video editing features.
8. The process of editing was very different in people of tier 2 and 3 cities. They engaged themselves in lots of neo formats like collages, adding non existing elements to photos like captions, emoji, filters etc.
9. The usage of camera as a documentation tool was extensive. People captured their bills, invoices, to-do lists, interesting patterns etc through their cameras. The use of cameras to capture information about events, phone numbers is extensive.

Proposed new AI features

The AI shot with basic edits

With such a huge pool of edits from the gallery, there is enough data on the kind of edits that are made on different types of photos. This data can be used to generate and automatically cropped, level adjusted photos that are captured as the user wishes, instead of being editing post-capture.

The video below illustrates this.

The edits are done in the camera itself to avoid redundancy and the piling up of images in the gallery, which came up as a problem from multiple users.

But even if a user chooses not to use the AI camera, the edits can be accessed through the gallery. Refer to the video below:

AI Shot: Documents

Similarly, the camera can recognise documents that are being captured and edit them accordingly for crop, brightness and contrast. There will also be an option to convert to PDF. Refer to the video below:

AI Shot: reminders, links and numbers

Many applications use the camera. Moreover, the camera actually forms the starting point to many of these journeys. For instance, taking a picture of an event that is about to happen, or maybe taking a picture of someone’s visiting card or a phone number that appears on some poster. These are all use cases that start with the camera. Or capturing important information that is written or printed somewhere.

All these journeys form interesting applications for an AI camera. Like for instance, take a photo of an event and set a reminder for the same automatically etc.

A video for this concept will be coming up soon.

Epilogue

It was great fun to hear the multitude of stories that came our way during the user research. It was fascinating to hear about the camera being used in so many unusual ways; especially from people in the above 40 age group and in tier 2 and 3 cities. Prototyping interactions for the camera was also great fun. This project was done in conjunction with Roshni Chatterjee.