Skip to content

Product updates

Learn Vocabulary from Photos: Words for What You See

Can a photo teach you words? See what research says about pictures and vocabulary, and how Radicard now finds words in photos that contain no text at all.

By iranimij Published 6 min read

Yes, you can learn vocabulary from photos, even photos with no writing in them. Every room and street is full of objects you can see but cannot yet name in your target language. This guide covers what research says about pictures and words, and how Radicard (version 1.8.4, November 2026) now picks out vocabulary for the things it sees when a photo has no readable text.

Why learn vocabulary from the things around you?

Because the objects in your own life are the words you will actually use. A fridge, a kitchen counter or a street corner contains dozens of nouns, and they are far more relevant to you than a list from a textbook chapter.

Everyday words also tend to come with a ready-made memory hook: you can look up from your phone and see the thing itself. That is a different kind of input from reading a sign. If you have not tried it yet, our guide to how to remember vocabulary explains why attaching a word to something concrete helps.

Are pictures better than translations for learning words?

Not reliably. The research is mixed, and it matters how you use the picture. In classic memory research, pictures are often remembered better than words, which is called the picture superiority effect. Nelson, Reed and Walling (1976) studied it in paired-associate learning, and later work showed it depends on how distinct the pictures are.

For second-language words, the evidence is less clear-cut:

  • Lotto and de Groot (1998) taught 56 Dutch adults 80 Italian words. Learning with translations produced better performance than learning with pictures (Language Learning, 48(1), 31–69).
  • Carpenter and Olson (2012) examined whether pictures help foreign vocabulary and found that learners tend to be overconfident about what they gain from pictures (Carpenter & Olson, 2012).
  • Zhang, Sala and Gobet (2025) ran a meta-analysis of the field. They compared pictures only, first-language words only, and the two combined, and concluded that combining both has benefits (Zhang et al., 2025).

The practical reading: a picture alone is a weak card. A picture plus the word plus a translation in your own language is a stronger one.

What is the difference between a photo with text and a photo without?

A photo with text gives you the words someone wrote; a photo without text gives you the words you would have to say. A café menu or a train sign already contains the vocabulary, and you only need to pick the useful words. See sentence mining for how to learn words in the sentences where you found them.

A photo of a kitchen has no written words at all. The vocabulary is in the objects: kettle, cutting board, sponge, fridge magnet. This is the kind of everyday vocabulary that courses often skip and that you need the moment you try to talk about your own home.

How does Radicard find words in photos without text?

If a photo has no readable text, Radicard picks out vocabulary for the things it can see. This arrived in version 1.8.4 in November 2026.

Here is how it works:

  1. Open Radicard and photograph a scene: a kitchen, a street, the inside of a fridge.
  2. Radicard looks for readable text. If there is none, it looks at what is in the photo instead.
  3. It picks out vocabulary for the things it can see, at your level (A1 to C2), with translations.
  4. You save the words to your on-device library, like any other scan.

For example, a photo of an open fridge might give a beginner words for milk, eggs and butter, in the language you are learning, with translations.

Radicard is a vocabulary app: you photograph any text and Google’s Gemini AI picks out 5–15 words matched to your CEFR level, the European scale for language skill. Saved words are reviewed with the Leitner system, seven boxes with review gaps of 1, 2, 4, 8, 16, 31 and 120 days. See how it works, try the web app, or get the app from the App Store or Google Play.

Which scenes give you the most useful words?

Cluttered, everyday scenes work best, because they hold many nameable objects at once. A tidy empty wall gives you almost nothing, while a busy kitchen counter gives you a dozen words in one shot.

SceneTypical words you can expectGood for
Kitchen counterkettle, sponge, cutting board, jarCooking and home talk
Open fridgemilk, eggs, butter, leftoversShopping and meals
Street cornerbench, traffic light, bicycle, binGetting around
Your deskcharger, notebook, mug, lampWork and study talk

The table lists examples, not a guarantee of what any photo will return. Take the photo in good light, from close enough that objects are clear, and favor scenes you meet every day. Words you see daily get revisited without effort, which supports your review schedule.

What does photo vocabulary not do?

A photo gives you words to learn; it does not teach them by itself. Four honest limits:

  • A scan needs internet. The image is read by Google’s Gemini AI, so scanning a new photo is online-only. Reviewing saved words works offline.
  • The free plan has 5 scans a day. Premium (€39 a year, with a 7-day free trial) is effectively unlimited.
  • Pictures can make you overconfident. Seeing the object feels like knowing the word. Test yourself on it later.
  • You still need to check. A word offered for an object is only as good as the picture. If something looks wrong, skip it.

None of these studies tested photo scanning by an AI app. The research above is about pictures as learning aids in general, so treat this as a sensible method rather than a proven one.

How should you review words from a photo?

Review the word without the photo, then check the photo. That is retrieval practice, which means pulling a word out of memory rather than re-reading it. Cover the translation, say the word aloud, and only then look.

Two habits help:

What should you do next?

Start with one room today. Five steps:

  1. Pick one place you know well: your kitchen, your desk, your street.
  2. Photograph it with Radicard, or list ten objects you cannot name in your target language.
  3. Save the words and say each one aloud while looking at the real object.
  4. Tomorrow, test yourself on them without the photo.
  5. Add one new place each day, and mix in text photos such as menus and signs, as we describe in comprehensible input and flashcards.

Frequently asked questions

Can you learn vocabulary from photos?

Yes, as a source of words and a memory cue. A photo of your kitchen or street shows you the things you could name in your target language but cannot yet. Research is mixed on whether pictures beat translations, so pair each picture with the word and a translation, then review the word later.

Are pictures better than translations for learning words?

Not reliably. Early lab work by Lotto and de Groot (1998) found that learning with translations produced better performance than learning with pictures. A 2025 meta-analysis by Zhang, Sala and Gobet instead points to combining first-language words with pictures as the approach with benefits.

What does Radicard do with a photo that has no text?

Since version 1.8.4 (November 2026), Radicard picks out vocabulary for the things it can see in the photo, such as items in a kitchen, a street or the inside of a fridge. The words match your level and come with translations, and you save them to your library like any other scanned words.

Do photo-based words need an internet connection?

Scanning a new photo needs an internet connection, because Google’s Gemini AI reads the image. Studying the words you have already saved works offline. The free plan includes 5 scans per day.